Report

INTRODUCTION

Throughout the 1980s there was growing criticism of the 1976 Fair Employment Act and, to some extent, of the Fair Employment Agency (see, for example, articles in Fortnight Magazine Nos. 124 2/4/76 Pp6-8; 178, Oct./Nov., 1980, Pp8-9; 188, Oct., 1982. Pp8-11; 243, Sept.9 1986s p13). Some of this criticism focussed on weaknesses in the Act and argued the need for strengthened legislation (eg McCrudden, 1981,1983; SACHR, 1987) while others questioned the commitment of the Government to fair employment and attacked the approach adopted by the FEA (eg Graham, 1984; Rolston, 1983; Rolston and Tomlinson, 1988).

From the mid-1980s onwards a series of pressures led to the passage of a new Fair Employment Act in 1989. Three predominant Pressures can be identified: firstly, Government statistics indicated that the unemployment gap between Protestants and Catholics in Northern Ireland was not declining (see sections eight to ten of this review); secondly, the Standing Advisory Commission for Human Rights carried out a wide-ranging review of the 1976 Act and recommended additional legislation (SACHR, 1987); and thirdly, external Pressure, Particularly from the United States and the MacBride Campaign (McCormack and O'Hara, 1990), was viewed by some as threatening both inward and existing investment in Northern Ireland.

This final section of the review will examine the measures incorporated into the 1989 Act and discuss some of the research issues that could develop over the next few years.


FAIR EMPLOYMENT (NORTHERN IRELAND) ACT 1989

The 1989 Act introduced a number of new measures in comparison with the 1976 Act. The FEA was replaced by the Fair Employment Commission (FEC) which had additional resources and Powers. Whereas the FEA had both considered and judged on individual cases of alleged discrimination, this function was now to be handled by a new body, the Fair Employment Tribunal (FET).

In addition to these structural changes, the 1989 Act introduced a number of other important changes (see Fair Employment Today, Spring, 1989):

Firstly, all employers with more than twenty-five employees were required to register with the FEC, monitor the religious composition of their workforces and submit annual monitoring returns to the FEC. Within two years this provision will apply to all employers with more than ten employees. All public authorities and all employers with more than 250 employees are required also to monitor the religious composition of applicants - other employers are encouraged to retain application forms.

Secondly, indirect discrimination was made illegal. Thus the Act made it illegal for an employer to use a criterion for recruitment or promotion that had the effect of disadvantaging members of one community over another, where that criterion was not job-related.

Thirdly, employers were required to monitor their employment practices at least once every three years in order to determine whether or not action was needed to achieve fair participation. Employers who did not meet their statutory obligations could be removed from the FEC's register and become ineligible for Government grants or for public sector tenders.

Fourthly, the Act permits employers to initiate affirmative action policies to deal with under-representation in their workforces. While employers were encouraged to initiate affirmative action measures voluntarily, they could be directed to carry out such action by the FEC. In the Code Of Practice which accompanied the Act (DED,

1989 Pp27-32), legal affirmative action was defined as follows:

Affirmative action means action designed to secure fair Participation in employment by members of the Protestant, or members of the Roman Catholic, community in Northern Ireland by means including:
- the adoption of practices encouraging such participation; and
- the modification or abandonment of practices that have or may have the effect of restricting or discouraging such participation.

The system of religious monitoring will provide a detailed source of information on the bulk of the employed population in Northern Ireland, with annual data allowing for an analysis of labour market processes over time. More particularly monitoring will provide information on the experience of Protestants and Catholics in the labour market and provide a mechanism whereby the impact of affirmative action programmes can be assessed. The additional resources available to the FEC should ensure that this information is subjected to rigorous analysis and it seems likely that, at some point, monitoring data will be made publicly available, thus allowing for external analysis.

The FEC might consider the use of 'equality ratios' as a method for disseminating monitoring results (Whitfield, 1989). Equality ratios have the advantage of producing

a single, easily understood measure of the extent to which an employer is affording equality of opportunity in the employment process. Because an equality ratio produces a single measure it also provides a relatively simple method for charting change over time.

Figure 11.1 provides an example of how equality ratios might be used, in this case for data on employment in the Northern Ireland Civil Service derived from monitoring reports produced by the NICS Equal Opportunity Unit (NICS Equal Opportunities Unit, 1986.,1987; 1989). In the NICS monitoring reports employees are categorised into eight grades, with Grade A the most senior and Grade H the most junior categories. Figure 11.1 shows data for three years, 1985, 1987 and 1989. For each year an equality ratio has been calculated for each grade level.

An equality ratio is derived by finding the relative rate of employment of members of two social groups, in this case Protestants and Catholics. Thus, a ratio of 1.0 would imply that equal numbers of Protestants and Catholics were being employed, while a ratio of less than 1.0 would imply that more Protestants than Catholics were being employed. Two conclusions are evident from figure 11.1. Firstly, it is clear that while relatively large numbers of Catholics were employed in the more junior grades of the NICS, relatively few were employed at the more senior grades. Secondly, across the five years covered by these data there was an increase in the employment of Catholics, especially from grade F up to grade C; however, there was no clear pattern of improvement at the two most senior grade levels. Both conclusions are in accord with the most recent assessment of equal opportunities in the NICS (Osborne, 1990).

Figure 11.1 is included as an example of the way equality ratios can be used to illustrate patterns of employment between two groups both across grade levels and across time. In this case the data were based on those employed in the NICS although it would be possible also to derive equality ratios for recruitment or promotion figures.

The monitoring system introduced by the 1989 Act will not, however, produce a complete picture of the labour market in Northern Ireland. On the one hand the monitoring information excludes a number of groups, including those working in small firms and part-time workers, ie those working less than sixteen hours per week. An FEA investigation into employment patterns in the (London)Derry area (FEA, 1983) suggested that employment in small firms may play an important role in the assessment of the labour market in a particular area. Similarly, the exclusion of part-time workers from monitoring specifically excludes many women workers. The total number of excluded workers will amount only to a minority of the total employed workforce, but it will place limitations on what can be deduced about the operations of the labour market overall. A further limitation is that multi-site companies will file only a single monitoring return: this will allow for analysis of the labour market at a province-wide level, but inhibit analysis at sub-regional levels (it should be noted that the FEC can request monitoring information for each site: DED, 1989, p17).

Having pointed to some weaknesses in the monitoring data itself, we can now go on to consider a number of potential difficulties in the analysis of monitoring data. The first difficulty lies in the fact that the availability of data need not imply that all analysts will reach the same conclusions regarding its interpretation. This can be best illustrated by an example. In 1985 the FEA published an investigation of the Northern Ireland Civil Service. The FEA concluded that the proportion of Catholics in the NICS had been very low before 1968, but had increased substantially in the 1970s. Although Catholics were not adequately represented at the key policy-making levels of the NICS, the FEA report contained analyses which suggested that Catholics were moving up the promotional ladder. However, these conclusions were criticised in a series of articles (Miller, 1985, 1988). Miller, who had been a consultant to the FEA on the NICS investigation, claimed that analyses he and others had carried out on the data had pointed to a quite different conclusion, ie that "the imbalances in favour of Protestants could persist well into the next century" (Miller, 1988s p4).

The terms of this debate -had very technical aspects, including the relative merits of log-linear versus analysis of variance techniques, but it is included here merely to illustrate the point that different analysts can arrive at different conclusions despite using the same dataset. It should only be added that, following the FEA investigation, the NICS established an Equal Opportunities Unit which has produced three monitoring reports (data from which was used above, cf. figure 11. 1) allowing for further examination of employment and recruitment patterns in the NICS (Osborne, 1990; although see McCusker, 1985).,In addition, the Equal Opportunities Unit has been willing to bring problems identified by its monitoring into the public domain (see, for example, Harbison, 1990).

A further difficulty lies in the limited information on religious affiliation, population and labour availability in local areas, against which monitoring returns can be compared. In the United States a 'total availability estimate' is used which includes such factors as minority population, minority unemployment, minorities with requisite skills and availability for training (Whitfield, 1989, p29). By contrast, in Northern Ireland only a limited number of published Government social statistics (eg census information, CHS data) are routinely disaggregated by religion. Given this relative lack of baseline comparative information in Northern Ireland it may be that the monitoring data itself could be used to compare religious patterns in workforces in particular areas. There is, however, another view that labour availability data may not be appropriate insofar as they describe a situation which is a reflection of past discrimination (Edwards, 1990, p25).

A final point concerns those people who, by virtue of being unemployed, will not appear in monitoring returns. Osborne and Cormack's (1987) analysis of the 1981 census (see section seven) indicated that differential unemployment levels among Protestants and Catholics represented the main socioeconomic difference between the two communities: if the socioeconomic distribution of the employed population only was examined, the distributions for Protestants and Catholics were very similar. Because monitoring data relate to the employed population only, a similar pattern might be expected when Protestant and Catholic distributions are compared at a broad level of analysis. It will be important to set any such patterns within the context of the total labour market and, in particular, comparative unemployment levels (issues relating to unemployment are discussed further below).


AFFIRMATIVE ACTION

The FEC will have wider powers to issue directions to employers to implement fair employment policies. The simple fact of having global monitoring returns will, in itself, provide a strong base for action on the part of the FEC.

The Act closes what is involved in affirmative action measures and it is likely that few employers will be in a position to prevaricate about the necessity of such measures. Indeed, the fact that employers will have to collate annual monitoring returns makes it likely that many will delegate specific responsibility for fair employment concerns. It is likely also that the biggest employers will establish equal opportunities units similar to that of the NICS.

There exists an extensive literature on affirmative action measures and in the framing of the 1976 and 1989 Acts much attention was paid to practice in other countries, most notably the United States, Canada and New Zealand (see, for example, Glazer, 1975; Little and Robbins, 1982; Polenberg, 1980; Bindman, 1981; Dworkin, 1981; Ziller, 1984; EOR, 1986; Edwards, 1987; Mullan, 1988; McCrudden, 1986a, 1986b, 1988; Osborne and Cormack, 1989; Cormack and Osborne, 1989; Cross, 1990). Discussions on this issue are often clouded by conceptual confusion (Edwards, 1987, p 14); indeed, Jencks (1988) has suggested that US politicians happily endorse the concept of equal opportunity precisely because it means so many different things to different people. Arguably, in the framing of the 1989 Act greater attention was paid to the Canadian example of employment equity than to US practice, largely because of opposition to quotas (Cormack and Osborne, 1989; Osborne and Cormack, 1989), although ironically, recent Supreme Court decisions in the US have marked a retreat from the liberal policies of the 1970s and early 1980s (see EOR No.28, 1989, Pp23-25).

In Northern Ireland a further polemical edge was given to debates because of the campaign in the United States over the MacBride Principles (EOR No.8, 1986, Pp 1 6-21; Rolston and Tomlinson, 1988, p63; McCormack and O'Hara, 1990, Pp36-39). For the present purposes the most significant element of affirmative action, as presently defined in Northern Ireland, is that it excludes the use of preference at the point of selection, or quotas. Quotas have been opposed on a number of grounds, including that they are "restrictive, inefficient, exclusive and unjust" (DED, 1989, p31) and that they should not be used until other measures have been tried (SACHR, 1987).

Lawful affirmative action measures would include targeting training in such a way that under-represented groups would benefit, encouraging applications from under-represented groups or agreed redundancy schemes that avoid problems created by the 'last in, first out' rule. It would be lawful to establish criteria for such measures that were over-inclusive of under-represented groups, but unlawful to establish criteria that exclude over-represented groups. For example, it would be illegal to select only Catholics for training, but it would be acceptable to locate training facilities in a predominantly Catholic area or to select those who are unemployed for training. These measures seem quite strong although it has been suggested that they may contradict the 'merit principle' and could run into problems in the courts (EOR No27, 1989, Pp27-29).

One effect of the above measures may be that discussions on the links between religion and employment become more widespread and open than before. Undoubtedly many of these discussions will be marked by heated polemic and controversy, although even this is probably preferable to reticence or avoidance.


FAIR EMPLOYMENT TRIBUNAL

A further set of issues relate to the Fair Employment Tribunal. Undoubtedly the decisions of the Tribunal will be crucial in establishing the burden of proof necessary in claims of direct and indirect discrimination. It remains to be seen both just how rigid these criteria will be and how long it will take for them to be established (for an account of the extent of case law on discrimination in Britain see Carby-Hall, 1987).

At the time of writing an embarrassing loophole in the 1989 Act arose when the Tribunal ruled that an employer could not reveal the religion of a successful applicant, even though this basic information was necessary in order to determine whether or not there were any grounds for a complaint of individual discrimination. The net effect of this ruling has been to bring to a halt all cases alleging individual discrimination (see Business Telegraph [Belfast Telegraph supplement], 16/10/90, p3), although it seems likely that the Government will move to close this loophole.

A related point is that employers can challenge directions from the FEC by appealing to the Tribunal: once again a lot will depend on the judgement and speed of the Tribunal. An issue that may arise in this situation concerns the establishment of a baseline for judging whether or not a particular workforce is imbalanced. In FEA investigations judgements were often based on the 1981 census, although with the passage of time this becomes less satisfactory and it will be some time before data from the 1991 census are available. In the interim the FEC may have to come up with a system that is acceptable to the Tribunal.

One solution would be for Government social statistics to be routinely disaggregated by religion, thus allowing for the calculation of availability statistics for those entering, leaving and moving within the labour market. An alternative would be to use the monitoring data itself to provide baseline figures. In this way, the monitoring data for all employers within a specified area could be compared, with initial attention being directed at those employers where the religious composition of the workforce differs considerably from the general local pattern. Of course, this method runs into the objection already noted above that the general pattern in an area may in itself reflect past discrimination. A third alternative is to concentrate on patterns of recruitment over time and not worry about determining a baseline measure (Whitfield, 1989). Whitfield does not deny that low rates of applications from minority group members are a problem in equal opportunity measures, but, he argues, if an employer is seen to be applying equal opportunity in short-listing and recruitment practice, then the number of applications from minority groups will rise.


FORMAL PROCEDURES IN EMPLOYMENT PRACTICE

Both the 1976 and 1989 Acts emphasised the importance of fair employment procedures: to that extent that concentrate on structural solutions which seek to formalise recruitment and promotion procedures. However, there is a view that this presents an overly simplistic view of labour market processes. Maguire (1986, 19889 1990), for example, suggests that informal recruitment methods are not adopted for reasons of convenience or habit, but rather provide a means of controlling the workforce (see section ten), while Jewson and Mason (1986) point to the way informal methods of redeployment within large companies can effect employment status while avoiding formal promotion procedures. In addition, Jewson and Mason illustrate how formal promotion procedures can he subverted through tailored job specifications. More recently, Jewson et al. (1990), reporting on a study of employment practices and ethnic minorities in six organisations in Britain, concluded as follows:

Our case studies do not suggest that there is a single formulator employment practice which can be said to guarantee the fair and equal operation of procedures or equality of distribution of members of different ethnic categories in the workforce. The effects of formalisation are equivocal. The existence of formal procedures in itself is no guarantee of fairness. The principles on which formalisation is built may themselves be unfair. Even when fair, formal procedure may not be implemented. Even when implemented it may be subject to manipulation and circumvention. It may, moreover, be merely token (Jewson et al., 1990 p 12).

A further issue concerns the unemployed who, by virtue of their status, will not appear on any monitoring return. As the discussion in section ten showed, much of the research in recent years has dealt specifically with this group (see also below). The unemployed form a large segment of the employable population, but the full extent of the group is obscured by variability in the definition of who they are and changes in the regulations regarding the official unemployment count. Here again bringing the unemployed into the picture is necessary if a complete understanding of labour market processes is to be gained.

The question of unemployment has, however, a more fundamental importance for the 1989 Fair Employment Act. In the discussions prior to the 1989 Act most attention focused on the unemployment gap between Protestants and Catholics. Indeed, it was often taken as one of the failures of the 1976 Act that the unemployment gap was of the same order in 1985 as it had been in 1971. In other words, the continuing significance of the unemployment gap was to be one of the most important grounds upon which the need for further legislation was based. However, during the debates on the Bill in Westminster, particularly in discussions of the five year review of the legislation, Government ministers attempted to steer attention away from the unemployment gap as a criterion of the legislation's success. Instead they talked of the need for "the broadest possible approach" which took "account of all the factors influencing access to employment, including job creation and promotion". In other words, the suggestion appears to be that the criteria upon which the legislation will be judged will be somewhat more specific, perhaps based on data derived from the monitoring information, than the global measure of labour market activity provided by unemployment levels disaggregated by religion.

Part of the argument for not using unemployment levels to assess the impact of the 1989 Act is based on the idea that, as a global measure, unemployment levels provide only a very crude estimate of labour market activity. Thus, for example, a study into the relationship between employment and unemployment in Britain (Dicks and Hatch, 1989) suggested that in order to lower the unemployment total by one person, two jobs had to be created. The reason for this is that recruits to jobs do not come only from the unemployed: in a situation where jobs become available, for example, potential recruits might come from married women, who would not ordinarily be registered unemployed. Arguably this relationship between employment and unemployment is exacerbated by changes to the method of counting unemployment (Dicks and Hatch, p29)

In Northern Ireland the situation is complicated further: Gudgin and Roper (1990) have calculated that in order to equalise unemployment levels in Northern Ireland and the UK by 1995 there would have to be created either an extra 163,000 jobs, if they were predominantly in the manufacturing sector, or an extra 188,000 jobs, if they were a mix of manufacturing and service sector jobs.

The implication is that, in general, to remove one person from the unemployment register around three additional jobs need to be created. (Gudgin and Roper, 1990, p69).

The argument regarding the assessment of the impact of the 1989 Fair Employment Act might then be presented as follows: we know that there is an uneven relationship between additional employment and levels of unemployment generally, and that this relationship is more uneven in the particular case of Northern Ireland. Therefore, in order to assess the impact of the 1989 Act it is more accurate to use targeted measures, such as the Protestant/Catholic ratios in recruitment or promotion, rather than gross measures such as that provided by unemployment levels. Furthermore, unemployment will only come down with massive job creation, but the purpose of the 1989 Act is to impact on recruitment, promotion and redundancy practice, not job creation: better that the assessment of the Act concentrate on the first three, rather than the last of these.

It is up to the reader to judge the merits of this line of argument, but, if only from the point of view of political appearance, it seems unlikely that the unemployment gap will disappear from discussions of the, impact of the 1989 Act. if, for example, in five years time the level of Catholic male unemployment is still two-and-a-half times that of Protestant male unemployment, it may be difficult for the Government to resist pressure to introduce strengthened legislation.

There is a further argument against placing undue emphasis on unemployment differentials, although it takes quite a different form from those arguments considered above. Brown (1990) and Robinson (1990) examined unemployment differentials between whites and minority ethnic groups in Britain. Both analyses showed that, in periods of rising employment, the unemployment gap between whites and minority ethnic groups narrowed, only to reappear when unemployment started to increase again:

The lessons are that good news about unemployment is not necessarily good news about equality, and that getting a job is not the same as getting a good job
(Brown, 1990, p48).

Although the Fair Employment Agency was subject to much criticism, one undoubted value it had was the amount of research information it generated, either directly or indirectly: the extent to which FEA-sponsored research has featured in this review illustrates this point. The 1989 Act creates a new context for research and it seems likely that the FEC will play a particularly important role in setting the research agenda for the next few years, not least because of the monitoring data. In addition, further detailed analyses of influences in and on the labour market would be possible if more Government social statistics were routinely disaggregatcd by religion. If this were to happen, and in the context of religious monitoring of employees the case for routine disaggregation of Government social data by religion seems overwhelming, one might expect a steady stream of research reports on the relationship between religion and employment in Northern Ireland over the next few years. It is possible, however, to point to a number of research areas that may deserve particular attention.

Section five of this review Pointed to the value of social mobility surveys using data from the turn of the century, the 1950s and the 1970s. There may be value in carrying out a further study of social mobility in Northern Ireland in the 1990s to assess whether the common patterns identified in the earlier studies have changed.

Section one of this review provided a simplified model of the employment process and suggested that most existing research dealt with the early stages of this process. By contrast there appear to be few studies which examine processes at the point of selection, where subjective processes play a potentially larger role(although see Kremer et al., 1989). This is an area that may benefit from further attention.

Barring a number of exceptions (see sections nine and ten) there appear to be relatively few published studies of processes within particular workplaces in Northern Ireland. Studies of this kind could contribute to our understanding of the 'chill factor' or intimidation, and would throw light on the way in which legal measures are actually implemented in practice. The particular lack of such monarch in Northern Ireland is unfortunate.

Although there is a great deal of research on educational issues (see section four) there is a strong case for additional work on vocational and further education, and on training for employment. Two general issues suggest themselves as important: firstly, the provision and uptake of places in vocational education and training; and secondly, an examination of the processes mediating the move from vocational education and training into employment. Religious monitoring of vocational education and training would provide a useful starting point for research in this area.

Following the passage of the 1989 Act the issue of fair employment moved from the centre of political debate. There is every reason to believe that this situation is temporary, however: it might be expected that as soon as the first results from the 1990 monitoring exercise become available the issue will once again return to the centre of political and academic debate. This issue of the Majority Minority Review provides an informed background fist those debates and discussions; a future issue will provide an account of the debates and discussions yet to come.


NOTES
Figure 11.1 is based on data in NICS Equal opportunities Unit (1986,1987,1989)