Prices and Incomes Accord, Labour Market Analysis and Employment Planning, Working
Paper No.37, May 1990, p.19.
31 D W Oxnam, The Incidence of Strikes in Australia', in J E Isaac & G W Ford (Eds.),
Australian Labour Relations Readings, Melbourne, 1966, p.18.
be traced primarily to a particular issue, they say, "analysis would require more than just the seven categories used by the Australian Bureau of Statistics".33 As Dabscheck and Niland point out, the decline in one category of stated cause as a problem area leads to an automatic increase in other categories, even when there may have been little or no increases in disputes in those areas.34
Several limitations in the use of ABS statistics have also been noted by Plowman, Deery and Fisher. They itemise their reservations thus: the limit of ten working days lost may well understate the incidence of strikes because those of short duration involving small establishments are not included: statistics relate only to establishments where stoppages actually occur; effects such as unemployment and production lost in other associated places are not recorded: the industry classifications are not detailed enough; strike statistics include lockouts which could give a distorted view of worker militancy or intransigence; subjectivity enters into the collection and classification of data so that marked variations occur, (for example ABS and NSW Department of Labour Statistics); and, strikes may be multi-causal yet are recorded as having only one cause.35 The last two criticisms were noted earlier by Oxnam.
Problems can also arise from competing sections within the one recording agency. Don Rawson in his study on trade union membership outlines the discrepancies which exist between the annual series of Trade Union Statistics’ and another ABS series called Trade Union Members' which were produced in 1976, 1982, 1986 and 1 9 8 8 .3 6 Although the second series provides valuable information about the composition of unions since 1976, Rawson points out that they also give a different impression on union density from the first series.37 The older Trade Union Statistics’ suggest that union membership has remained fairly static at around 55- 56 per cent, while those from the Trade Union Members' series point to a decline in union density over the same period. Rawson concludes it is legitimate to say that both series have value, however, "one cannot escape an uneasy feeling that each casts some doubt on the validity of the other".38
33 B Dabscheck & J Niland, Industrial Relations in Australia, North Sydney, 1981, p.74.
34 Loc.cit.
35 D Plowman, et al, op.cit., p.45.
36 ABS Catalogue Nos. 6230.0 and 6325.0.
37 D Rawson,(a) Is Unionism Everywhere In Decline, paper presented at Australian Political Studies Association Conference, University of New England, Armidale, 26-28.8.88
Ironically perhaps, no criticism of ABS figures is registered in the Report of the Committee of Review into Australian Industrial Relations Law and Systems
(hereinafter called The Hancock Report). The aim and purpose of the Committee was to develop a more effective and practical industrial relations system. To this end the Committee was given broad terms of reference, which included specifically the direction to examine, report and make recommendations on "all aspects of Commonwealth law relating to the prevention and settlement of industrial d is p u te s ".39 A review of the available evidence about industrial disputes in Australia was undertaken in regard to strikes and lockouts. "Because of their comprehensiveness and their availability for a long period of time", the Committee reported, "the ABS data are the most suited to our needs and we rely on them below".49 The tables of statistics which follow are accepted uncritically as the basis for the Committee's inquiry. There is no mention of bans as a feature of industrial disputation in Australia. This aspect of the Hancock Report will be discussed later within the context of the Report itself.
Beggs and Chapman tested the proposition that the Accord has had no independent effect on Australian strike activity. Using working days lost per unionist from strikes over the period 1959 to 1983, econometric models using a range of variables were used such as price inflation, profits, inventories and measures of labour demand. They concluded that the variables influencing strikes were changes in the Consumer Price Index, overtime, inventories and profits as a proportion of Gross Domestic Product, the job vacancy rate, and a number of political and institutional factors.41
An identical exercise was carried out by Chapman and Gruen for the period 1986 to 1989. Working days lost per unionist from strikes were predicted from the parameters of the earlier equations with the outcomes being compared with actual working days lost. The results showed that strike activity had decreased markedly from the beginning of 1983 in Australia in a way not explainable by changes in measured economic variables. Chapman and Gruen held that the difference between the projected and actual working days lost "can be interpreted reasonably as the
39 Commonwealth of Australia, Australian Industrial Relations Law and Systems. Report of
the Committee of Review, Vol.2, Canberra, 1985. p.1.
40 Ibid., p. 125.
41 Beggs and Chapman, cited by Chapman and Gruen, (b)op.cit., p.28. See also J W Duncan & W E J McCarthy, 'What is Happening to Strikes?', in New Society, 22, 526, Nov.2., 1972, pp.267-269.
Beggs and Chapman in their earlier study had also examined the proposition that decreases in Australian strike activity were part of an international phenomenon and, by implication, unrelated to changes in the Australian institutional environment. Using a model in which annual strike data from thirteen OECD countries over the period 1964 to 1985 were used, Beggs and Chapman found the Australian experience of decreases in strike activity to be unique in an international context at that time 43 Chapman and Gruen further tested the model and results using data from 1964 to 1987. In brief they concluded that:
• decreases in Australian strike activity after the beginning of 1983 are most unusual; and
• that while decreases in strike activity are a world-wide phenomenon for the 1983-87 period, the fall for the rest of the world is about 40 per cent while the Australian diminution is around 70 per cent.44
Aside from the conclusions that can be drawn on the uniqueness of the Australian experience over the Accord years, these results have considerable relevance to this thesis. They show that the inclusion of data on working days lost from strikes may not have made any significant change to the research results. It is clear that both sets of data, that is on working days lost and on the incidence of industrial disputation, show a marked change in the pattern of activity during the Accord years. In this context it is worth mentioning the three caveats that Chapman and Gruen note in relation to their conclusions.
The first is that the models provide a very simple representation of influences on strike activity. The results therefore are indicative only.
The second is that conflicts between employers and employees can manifest themselves in less covert activity than work stoppages and that bans, go-slows and work to rules may have become different since the beginning of the Accord. They cite empirical evidence that there has been a change since the early 1980s in the form of industrial disputation towards the use of bans.43
42 Ibid., p.29. 43 Ibid., p.30. 44 Ibid., p.32.
45 Loc.cit. They are referring here to the research that was undertaken for this thesis and