THIS article reviews certain empirical results arising from the first phase of an investigation into the manual worker industrial wages structure. The major study will in part draw upon the material presented here in attempting a detailed analysis of earnings determination at the broad industry level. The aim of the present paper is, however, of more modest dimension. It is concerned with wage composition and the principal focus of attention will relate to the 'importance' of basic pay in weekly earning, 1968-74. The study proceeds as follows: first, the wider context of the major study is introduced simply to locate the largely descriptive approach here adopted; secondly, the nature of basic rate data is considered; thirdly, the principal data input of the present study, namely New Earnings Survey (N.E.S.) material, is introduced and analysed, and, fourthly, certain tentative 'explanations' of the observed relationships are advanced.
THE WIDER CONTEXTAlmost all post-Phillips studies of aggregate wage inflation in the U.K. have used wage rates (or, more accurately, W) as the dependent variable (Lipsey, 1960;Archibald, 1969;Lipsey and Parkin, 1970;Parkin, 1970; Thomas and Stoney, 197 1). This despite the widely observed postwar phenomenon of wage/earnings drift' (TurnerOffice of Manpower Economics, 1973) and clear evidence to the effect that changes in earnings may occur quite independently of changes in wage rates (Routh, 1959;Taylor, 1972). And, moreover, neatly side-stepping the view of many industrial relations theorists of institutionalist persuasion that the British industrial relations system has decomposed into two separate systems (Royal Commission on Trade Unions and Employers' Associations, 1968).The reliance of the economist upon a composite (male and female) index of wage rates is partly to be explained by data availability.* A more fundamental reason is, of course, that economics has no theory cf the rates-earnings-drift nexus, or a theory of the behaviour of bargaining structure which underlines this nexus. Consequently, the limited research within this area has relied upon ad hoc hypotheses as a source of testable predictions rather than conditional predictions drawn from some basic bedrock of economic analysis of wage compositionbargaining structure interaction (Paish, 1958;Turner, 1956;Dicks-Mireaux and Shepherd, 1962;Gillion, 1968).This relative lacuna in economic knowledge is paralleled by that attaching to the historical method of the institutionalist industrial relations approach, most clearly formulated in Donovan. Whilst the latter describes the way in which an informal, fragmented and autonomous process of bargaining at the plant level has grown up to occupy a place of at least equal importance alongside the formally recognised machinery of industry-wide bargaining it makes little or no attempt to 'explain' the phenomenon. Moreover, the institutionalist approach nowhere explicitly states the criteria which a good industrial relations system should satisfy. Consequently, unless one has esta...