Does individual relations policy affect productivity?

• Author: Peetz, David
• Position: Abstract

Abstract

This article considers the link between productivity, fairness, and industrial relations (IR) policy at workplace, national, and international levels using data from micro- and macro-level empirical studies as well as data from the Australian Bureau of Statistics (ABS), the OECD, and other sources. There is some evidence that policies that enhance fairness enhance economic performance. But the effects are conditional; they are neither consistent nor universal Government policies to encourage or discourage unions, to restrict the extent or scope of collective bargaining or related action, or to encourage or discourage non-unionism or individual contracting, will not do a great deal in net terms to improve economic performance. However, in any specific workplace, industrial relations and the decisions management makes can have a notable effect on productivity. While welfare and industrial relations systems do not make a large inherent difference to economic efficiency, they make a very large difference to social outcomes.

1. Introduction

   Athread through much economic policy discourse in the late 20th century was the alleged trade-off between equity and efficiency (Okun 1975). In the labour market, this is typically underpinned by the idea that the optimal allocation of resources will be achieved by the operation of a totally free labour market (Manning, 2004). Any divergence from that ideal in the interests of promoting 'equity' would be seen as harming efficiency. Yet the idea that efficiency and equity are opposed has been challenged by more recent developments in economics, which have led to the argument that more equal societies grow faster than less equal ones (Osberg 1995; Wilkinson and Pickett 2009).

   In debate on industrial relations (IR), in Australia and elsewhere, this conflict is manifested as arguments that policies should, and can, focus on productivity improvement, rather than equity. In recent times, a campaign to make fundamental changes to the Fair Work Act, on the grounds of improving productivity, has been pursued (Hewett 2011; Ridout 2011; Business Council of Australia 2012). This article questions the extent to which industrial relations policy affects productivity. Productivity is the quantity of output per unit of input. Labour productivity is output per hour worked. It is not measured by the value of that output, or the cost of that in put, or the amount of output not produced when there are no hours worked due to strikes. Debate is often complicated by confusion over the meaning of productivity.

   Proponents of particular IR policies often portray their preferred systems as being designed to enhance economic performance. The reason for that is straightforward. Almost everyone agrees that, other things being equal, people are better off in an economy with high productivity, high employment, and low inflation than the opposite. It is not possible to obtain the same sort of consensus about the distribution of income and power. So arguments about the allocation of resources and power will tend to be couched in terms of its benefits for the economy.

   IR policy often appears aimed at more objectives than it can meet. With few exceptions, it has much more of an impact in the long run on fairness, however defined, than on economic performance. If claims are made that a particular industrial relations policy is going to have very large (positive or negative) consequences for economic performance, such claims should be examined sceptically, as there is a reasonable probability that the effects may be small, even non-existent, or perhaps the opposite of what is claimed. The rest of this article considers the link between productivity, fairness, and IR policy at (in order) workplace, national, and international levels.

2. Micro-level evidence

   The major policy questions in IR focus around the extent to which policies advantage or disadvantage unionism, individual contracting or collective bargaining, and the taking of industrial action by unionised workers as part of collective bargaining, or the protection of employment. These are what contemporary debate on the Fair Work Act, and much of the debate on WorkChoices, has been about.

   There is a long history of studies in Australia and especially internationally that looked at the impact of unions on economic performance. There is a much smaller group of studies that look specifically at individual contracting.

   First, we refer to the studies on union effects. The ways in which unions can impede economic performance of a firm are by imposing restrictive work practices or by impeding the introduction of innovations such as new technology. I set aside the question of defining just what a restrictive work practice is. (Is it something that tempers unfettered managerial prerogative or a practice that management was willing to accept in the past but which it is no longer willing to accept?) There is some international evidence from the 1970s showing that restrictive practices had harmful effects (Elbaum and Wilkinson 1979; Lazonick 1979; Pencavel 1977). Such practices were common in Australia in that period up until the mid and late 1980s, but were mostly removed by the two-tier wage system, and then award restructuring and nearly two decades of enterprise bargaining. Restrictive practices were typically associated with demarcations arising from multiple unionism, but union amalgamations, single bargaining units, and the processes mentioned above substantially diminished or ended the impact of demarcations. As to whether u n ions restrict the introduction of new technology, while there were some cases of this, the evidence even from the 1980s was that, in general, unions did not substantially restrict new technology (for example Batstone and Gourlay 1986; Daniel 1987 McLaughlin 1979; Nichols 1986, p. 232).

   Still, it was generally thought amongst conventional economists that unions had a negative impact on economic variables until the emergence in the 1980s of a new literature, based principally around Richard Freeman and James Medoff's book What Do Unions Do? (Freeman and Medoff 1984). This showed that unions could have a positive effect on productivity through two mechanisms. One was through what they called the union 'monopoly' effect: unions raise wages and the higher wages lead employers to invest in labour-saving technology. This leads to higher labour productivity--though not necessarily higher multi-factor productivity. The second mechanism was the 'voice' effect: employees express their voice through unions and this leads to lower covert conflict at work and to improved techniques of production. In non-union workplaces, dissatisfied workers leave, causing turnover costs for employers; in union workplaces, they stay and seek to change the problems they identify. There is a body of evidence collected over the years that shows benefits from employee voice for economic performance. Direct and indirect participation by employees in decision making--preferably in combination--on average lead to lower absenteeism, lower labour turnover, higher morale and employee satisfaction, and higher productivity, though this may be conditional upon favourable workplace and institutional circumstances (Jones and Svejnar 1982; Strauss 1992; Zwick 2004; Grimsrud and Kvinge 2006).

   Whether unionism increases productivity is really a question of how far these competing factors offset each other. It is an empirical question that is likely to produce different results at different times and in different places. After Freeman and Medoff's book ca me out, there was mixed evidence from the United States. Some were in support of their argument (Allen 1985; Ben-Net and Estrin 1986; Phipps and Sheen 1994) some were counter to it (Addison, John, and Barnett 1982; Drago and Wooden 1992). Initial British evidence was adverse (Edwards 1987), but by the 1990s negative productivity effects from unionism appeared to have disappeared (Addison, John, and Belfield 2004). There was consistent evidence that unions reduced quits and increased job tenure (Freeman 1980; Addison, John, and Belfield 2004).

   More recently, three studies in Australia published in the last decade provide some evidence to support Freeman and Medoff. A positive relationship was found between unionism and productivity at workplaces where unions are active (Wooden 2000, p. 173). Collective bargaining coverage was associated with higher levels of self-claimed productivity (Fry, Jarvis, and Loundes 2002). Firms with high rates of union membership were more productive than firms with no union members (Tseng and Wooden 2001). Another study from the 1990s showed that the intensity of collaboration between management and workers (through unions) had a positive effect on workplace performance (Alexander and Green 1992). More recently again, and in contrast, a consultant's report was commissioned to show that reform of the building industry achieved 10 percent productivity gains through reducing union influence (Econtech 2007). Its core data have since been discredited, as either false or subject to selective or inappropriate interpretation (Allan, Dungan, and Peetz 2010).

   Two decades after the publication of What Do Unions Do?, the general consensus amongst those who reviewed the literature was that there was no consistent relationship evident between unions and productivity, with a wide variety of results; but the direct impact of unions on productivity tended towards zero. The impact, it appears, depends on circumstances (Addison, John, and Belfield 2004; Hirsch 2004; Freeman 2005; Kaufman 2005). Overall, studies from Australia and internationally suggest that unionised workplaces with good union-management relations and high employee participation or involvement will probably have higher average productivity than non-union workplaces. However, for those with adversarial and non-participatory union-management...