Collective Agreements and Relative Earnings in UK Manufacturing in 1973

The primary purposes of this paper are to investigate hypothesis that differences in trade union wage coverage between industries are a significant source of variation in earnings between industries and to attempt to estimate any union-non-union earnings The stimulus for this paper was publication, in New Earnings Survey (NES) for 1973, of a unique series of data which, inter alia, indicate percentage of workers in each industry whose wages are subject to agreements of various kinds. Prior to publication of collective agreement (CA) data only recently available data on extent of trade union organization in United Kingdom were published data on aggregate trade union membership. At industry group level of disaggregation these data must be allocated to industries in an ad hoc way. The result is that only limited reliance can be placed on them. Pencavel's (1974) recent study of relative wage effects of trade unions, only one for United Kingdom, used such data and is therefore subject to limitations of those data. The NES data have quality that they directly nieasure extent of coverage of unlion1 rate. As a result measures of union and non-union wages in this study are different from those used in most other studies. The union wage here is not estimated wage paid to uinion members as estimated from union membership data. It is, instead, estimated wage negotiated by union and paid to union and some non-union workers as estimated from union wage coverage data. In many respects this is main feature of this study that distinguishes it from most others. In literature in this area it is usually observed that there are wage spillover effects from union to non-union sector, making concept of a union-non-union wage differential elusive in empirical analysis. (See for example Lewis, 1963; Pencavel, 1974.) Hence we would expect that differential estimated from our data would tend to be considerably higher than that estimated in most other studies. We shall call concept of union-nonunion differential used in this paper the coverage differential.