The determinants of union membership : evidence from two Australian surveys

This thesis is an econometric investigation of what determines an individual's probability of union membership. Following Olson (1965) it is argued that unions are fundamentally collective organisations that seek to advance collective goals. Therefore, it is stressed that union membership needs to be understood as a collective action problem in which individuals have an incentive to free ride the collective benefits of unionism. The empirical analysis commences with the estimation of a simple reduced form model of union membership. Similar models of Australian union membership have been estimated by Crockett and Hall (1987), Miller and Rummery (1989), Deery and DeCieri (1991), and Christie (1992). Chapter 3 surveys the findings of the previous studies and presents new evidence from the Class Structure of Australia and Issues in Multicultural Australia surveys. In particular, the first Australian estimates of the impact of workplace size on the individual's probability of union membership are reported. However, in the following chapters we show that the standard model masks important relationships that are revealed by more disaggregated models. Indeed, a major contribution of this thesis is the development of more detailed, "multiple-hurdle" models of union membership. In Chapter 4 we modify the simple model by specifying a two-stage model of unionised employment and union membership. Here union membership is modelled as a decision which is conditional on the individual being employed in a unionised workplace. One of the principal findings of this chapter is that once the availability of union membership is controlled, there is no significant difference in the propensities of private and public sector workers to join unions. Another issue explored in Chapter 4 is the appropriate interpretation of the relationship between wages and union membership. Some researchers argue that there is a wage premium associated with union membership, and that this wage premium should be treated as a determinant of the individual's probability of membership. However, we argue that, under normal circumstances, an individual cannot expect a wage gain simply by joining the relevant union. Indeed, our estimates indicate that there is no wage advantage associated with union membership in unionised establishments. However, we do fmd a significant wage premium associated with employment in a unionised establishment. There is some controversy in the literature about the impact compulsory union membership arrangements have on the overall level of union membership in Australia. In Chapter 5 we specify a two-stage model of compulsory and voluntary union membership. Using the estimates from a model of voluntary union membership, we fmd that closed shop arrangements have a substantial impact on the overall level of union membership. In particular, we estimate as many as 65 percent of compulsory union members would not join a union if membership was not a condition of their employment. Finally, in Chapter 6 we investigate the reasons why a smaller proportion of women join unions than men. We fmd that the male-female membership differential is largely the product of two factors: (i) that women are less likely than men to be employed in unionised establishments; and (ii) that women have a substantially lower rate of closed shop employment than men. We fmd little evidence to support the proposition that women have a weaker desire for union membership than men