rules of thumb and broad notions of fairness. It is this elasticity, and not monopsony, that explains the ability to absorb higher wage costs. Second, however, behaviour is institutionalized: many firms operate in labour markets circumscribed by ethnic exclusion, and managers and workers develop expectations within such specific contexts as to what is fair in the light of knowledge of other local firms and the state of the trade. Pay is not immune to market forces, but these forces are interpreted through structures that limit knowledge of the wider economy and that shape norms of fairness. Third, within the shared context of competitive markets, the elasticity of pay structures leaves space for choice and idiosyncrasy. Firms in the same market and geographically close to each other were found paying notably different wages. Some of these differences could be traced back to differences of product market context, so that a restaurant offering a ‘quality’ product might pay more than one in a more basic niche. Yet just how much more it needed to pay depended on the preferences of the manager, the history of the firm, and the degree of active bargaining by workers. Other studies (Cox, 2005; Holliday, 1995; Moule, 1998; Ram, 1994) have shown that workers engage in tacit bargaining over pay, hours, and work load, drawing on their labour market power and their own past experience. For example, Holliday (1995) makes a distinction between ‘core’ and ‘transient’ workers. ‘Core’ workers were adaptable, loyal, and able to ‘fit in’ with the prevailing pattern of social relations in the firm. These workers were regarded as ‘one of the family’, and exerted a degree of influence over work loads, promotion and even who was recruited. ‘Transient’ workers lacked this leeway, and rarely stayed with the firm for any length of time.
This kind of evidence helps us resolve puzzles in contemporary accounts. The standard view of legislation on such things as minimum wages is that this legislation has determinate effects or at least tendencies. The most commonly expected effect is that wages will rise. Other employment legislation, for example that which controls the right to dismiss, is expected to limit managerial decision-making freedom. A less commonly articulated argument is that there may be a stimulus to economic efficiency, because increased costs will ‘shock’ firms into improving labour utilization. Yet evidence suggests that effects are often weak (Arrowsmith et al., 2003; Edwards et al., 2004). Part of the explanation is wholly consistent with conventional views: legislation is often modest in its aims, and thus any effects are small. But how then do we explain situations where there was likely to be an effect, for example firms that have to raise their wages? The answer lies in the context of many such firms, notably the fact that pay structures are far less fixed and clear-cut than usual models of pay setting expect. External influences are thus moderated. The effects of legislation are also not necessarily direct. For example, it was found that care homes were affected by the national minimum wage even though many paid above it and were little affected by it directly. But they were constrained in the prices that they could charge and they found that the advent of the minimum made other jobs more attractive. They thus found staff recruitment and retention difficult. Similarly, it was possible to identify those conditions under which the efficiency-encouraging aspects of legislation were activated.
Developing causal accounts
Institutional analysis has thus offered a rich list of factors that might explain apparent anomalies. The forces of history, custom, ignorance, and chance may all help to explain patterns that conventional views cannot. It has been possible to generate intellectually coherent explanations, an achievement that should not be minimized when the limited amount of research in this vein, as compared to the massive