While the recent increases in income inequality are well-documented, considerable controversy still exists over which factors have been the most important causes of this trend.[2] However, one of the essential claims in the popular writing on globalisation is its supposed impact on inequality. The role played by globalisation on labour market outcomes and income inequality in developed countries has been a particularly fertile ground for research during a time when international trade liberalisation has progressed and concerns about imports from low-skill-abundant, less-developed countries (LDCs) have been prominent.[3]
At the outset, it should be noted that increased earnings dispersion in the developed economies does not appear to have been primarily the result of shifts in employment from manufacturing to services. Even for the countries that witnessed the greatest increases in earnings dispersion, the increased-earnings dispersion phenomenon has been observed within narrowly-defined industries across the entire economy. On the face of things, the latter observation seems to rule out the more hasty of the trade-related explanations. Trade economists have long argued that the natural framework for thinking about the long-run effect of trade on labour markets, at least from a maintained assumption of competitive markets, is the Stolper-Samuelson theorem and its various generalisations.[4] Simply stated, the implication of the theorem for skilled-labour-abundant developed economies is that a reduction in the relative price of unskilled-labour-intensive goods caused by more liberal trade with LDCs will lower the relative return to unskilled labour in the developed economies. The simplest trade model predicts that increased trade with countries like China and India would worsen the distribution of earnings in developed countries.
The trade-theoretic account of trade shocks as running from commodity-price changes to factor-price changes provides a compelling equilibrium mechanism and some useful rough empirical checks. Nevertheless, the aggregate professional consensus would seem to have settled on the conclusion that trade has a relatively small effect on the skill-premium, but that other factors (especially skill-biased technological change) are more important. Consistent with this consensus, Gaston (1998) finds that the declines in Australian manufacturing employment were barely affected by lower levels of trade protection.[5] Gaston also finds that Australian real earnings were extremely resilient in view of the (then) most recent recession and trade liberalisation. In addition, he finds that the adverse employment developments were, at most, weakly linked to this real wage resistance.[6]
Compounding the problems of inference, it is extremely difficult to unravel the distinct impacts of technical change and international trade on labour market outcomes. For example, globalisation may lower the costs of diffusing new technology and encourage capital for labour as well as skilled for unskilled labour substitution. More speculatively, the rate of technical progress may be an endogenous response to the need to maintain competitiveness in the global marketplace. The same type of argument can also be made about increasing global competition and institutional changes, such as de-unionisation and the decentralisation of wage bargaining, which are both features of many developed economies, including Australia.
Furthermore, since countries with similar standards of living and economic development generally have access to labour and capital of similar quality, it is quite likely that the magnitude and nature of any technical changes will also be similar. In fact, it has been argued that this must also be true for any changes on the demand-side, since European Union countries were also affected by import penetration from countries abundant in unskilled labour (Katz et al., 1995). Given this similarity in aggregate endowment, technology and shocks, it seems quite natural to investigate the institutional forces operating in each country to explain cross-country differences in the trends and structure of earnings dispersion. In the United Kingdom and the United States, de-unionisation has been a significant labour market development, in economies in which structures are already relatively decentralised (Katz, 1993). Naturally, these changes are not independent of growing international competition. During the 1990s, Australia implemented a number of labour market reforms to decentralise its traditionally centralised form of wage bargaining. The argument made by employer groups has been that such changes were inevitable because of the need to maintain international competitiveness (Borland, 1999).
Overall, the widespread concern with globalisation may have emerged as a result of changes that are obscured when standard economics methods are used to study the labour market effects of globalisation. The primary focus of the existing literature has been on the direct effects of globalisation — the direct effects of the flows of goods and factors of production on labour markets — and on not the indirect effects of globalisation.[7] But indirect effects also operate on the labour market, by transforming the structures that support one set of equilibria and inducing change in those equilibria. Because economic and political structures are related, changes in the relationship of a national economy to the global economy can produce profound changes in the political-economic arrangements of a country. In addition to affecting equilibrium wages and employment, such changes could well be unsettling in themselves.
The altered roles of labour unions and the welfare state provide obvious examples of the importance of institutional differences. Part of the support for relatively unskilled workers’ incomes comes from the mutually supporting institutions of unions and welfare state. That is, as a result of labour market institutions, such as labour unions, some workers receive a higher wage than other otherwise identical workers. A common finding is that measures of wage centralisation are generally negatively associated with wage dispersion (for example, Blau and Kahn, 1996). Likewise, higher rates of unionisation and collective bargaining tend to be associated with a lower incidence of low-paid employment and less earnings inequality.[8] In fact, the increases in inequality in recent years have coincided with more decentralised wage bargaining and de-unionisation.
It should be clear that globalisation could affect the union bargaining strength and workers’ incentive to unionise, with straightforward implications for equilibrium relative wages. There is now a sizable body of research examining the relationship between the institutional structure of the unionised sector of an economy (that is, the extent and centralisation of organisation) and various measures of macroeconomic performance. Countries with encompassing labour market institutions (that is, large unionised sectors with centralised bargaining) are characterised by: lower wage inequality (Rowthorn, 1992; OECD, 1997); lower unemployment (OECD, 1997); and higher growth (Rowthorn, 1992; Calmfors, 1993). The usual explanation involves the ability of centralised bargaining institutions to internalise negative wage externalities (Calmfors, 1993). That is, where strong sectoral unions pursue wage gains relative to some perceived market wage, resulting in cost-push inflation, reduced employment, lower growth and inter-sectoral inequality, the centralised union recognises these negative externalities and takes them into account in its bargaining. Thus, as unionisation has declined, there is some evidence that wage inequality has increased (Freeman, 1998).[9]
Increased inequality, and real deterioration in the labour market outcomes of unskilled workers, is also directly related to changes in demand for welfare-state provision. For example, it has been observed that despite increases in the dispersion of earned incomes that, in some countries at least, inequality in post-transfer and post-tax income inequality has not grown (Gottschalk and Smeeding, 1997). This suggests that political pressures have been brought to bear on the generosity of public transfers at a time when earned incomes have become more unequally distributed. From a political-economic perspective, the growing inequality of income could be associated with strong compositional effects on the demand for public insurance. In particular, it seems to be the case that the growing size and economic significance of sectors of the economy that pay higher wages for certain types of workers, could result in political pressures that lead to higher levels of transfer payments to disadvantaged workers. This could result from changes in the identity of the median voter or as an optimal response to increased income risk in an increasingly open economy (Alesina and Rodrik, 1994).
Some scholars argue that the increased mobility of capital not only erodes the tax base, reducing the state’s ability to fund welfare programs, but by shifting taxes onto labour, the capacity of the state to redistribute is reduced (Tanzi, 1995). For example, some European countries, in the face of increased international competition, have tried to reduce the ‘generosity’ of their social programs (Gaston and Nelson, 2004 and Gaston and Rajaguru, 2007a). In ways that are harder to quantify, but seem prima facie plausible, the decreasing cost of the exit option increases the relative power of business in policy-making (Huber and Stephens, 1998). Finally, it has been argued that globalisation increases the general credibility of orthodox (that is, market-oriented) policy advice, thus reducing the plausibility of arguments supporting welfare state expansion and enhancing the credibility of arguments in favour of welfare-state retrenchment (Krugman, 1999). In the popular consciousness, at least, coincident with the recent onset of globalisation has been a move towards privatisation, deregulation, neo-liberalism and ‘economic rationalism’. In the case of Australia, John Quiggin (1999, p.240) argues that ‘Increased inequality is the result of the neoliberal reform program as a whole. The role of globalisation per se has been overstated.’
Finally, it is being increasingly recognised that globalisation has important social and political dimensions in addition to the usual economic dimensions of primary interest to economists. For example, recent research finds that social integration contributed to de-unionisation in OECD countries, while economic globalisation mattered far less (Dreher and Gaston, 2007a). Although largely neglected in the economics literature, both political integration and social integration are likely to be important for income inequality. For example, in the absence of restrictions on capital mobility, a country is more likely to competitively lower taxes or offer subsidies to attract investment, the closer a potential host country’s culture is to that of a source country and the easier it is to exchange information. Lower taxes may also lower social standards and this is one channel through which the social dimension of globalisation may be important for income inequality. On the other hand, political integration may ameliorate a potential ‘race to the bottom’ which may be induced by economic globalisation. Hence, while economic globalisation may increase inequality, political globalisation could actually serve to reduce it.
In the debate about the consequences of globalisation it is important not to take an overly narrow perspective of globalisation as this may severely bias conclusions about the ‘true’ effects, direct and indirect, on labour markets and income inequality. In addition to the more standard supply-and-demand factors, inequality may be adversely affected by changes in labour market institutions, trade liberalising policies (often bundled with privatisation and deregulation measures as well as changes to social policies; see Lindert and Williamson, 1991) and a variety of non-economic factors which have simultaneously affected many economies.[10]