جهانی شدن، تحرک سرمایه و رقابت در مالیات: تئوری و شواهد برای کشورهای سازمان همکاری و توسعه

Are the predictions of tax competition theory wrong? While the tax competition literature predicts that taxes on income from capital decrease with increasing globalisation, past empirical studies on various data find contradicting evidence. By using different data and additional elements of economic theory, this paper aims to challenge the empirical contributions. For a panel of 14 OECD countries and for the period 1967–1996, we find that globalisation has indeed a negative and significant impact on corporate taxes. Furthermore, globalisation tends to raise labour taxes and social expenditures. As a consequence, the so-called “efficiency” and “compensation” hypotheses of globalisation are not competing, but rather, both appear to apply at the same time. Efficiency has an impact on the tax-mix, whereas compensation is provided through increased social expenditures.

Following the basic results of tax competition literature, capital taxation is negatively related to the degree of international capital mobility, whereas labour taxation relative to capital taxation is positively related to international integration of national economies. The theory also suggests that larger countries levy higher capital tax rates than smaller countries because the erosion of their tax base is smaller in per capita terms (cf. Bucovetsky, 1991 and Wilson, 1991). These predictions are derived from general equilibrium models in full accordance with microeconomic principles. It is thus surprising that most recent empirical studies obtain almost reverse results. For instance, in a panel regression of 15 OECD countries for the period 1976–1990, Garrett (1995) finds that a rising exposure to international trade, which is used as a proxy for financial liberalisation, leads to an increase in capital taxation. Referring to cross-country studies of economic growth, Quinn (1997) considers a broader range of 64 countries with annual data averaged over the years 1974–1989 and concludes that levels of corporate taxation are positively associated with financial liberalisation under a wide variety of different model specifications. These findings are supported by Swank (1998). In a panel regression for 17 industrialized countries (mainly OECD countries) for the period 1966–1993, he finds that three different measures of capital mobility are positively related to the proxy of corporate taxation. A closer look at these empirical results reveals possible sources of problems and deficiencies. For instance, the creation of a proxy for the dependent variable—the tax policy—is a major problem. Proxies commonly used in cross-country studies are the revenues from corporate taxation as a percentage of the GDP. This method is inadequate in several respects. We set out the reasons in Section 3. In addition, specifications of estimated equations can be enriched through economic theory. Rodrik (1997) avoids the data problem by using effective average capital and labour tax rates. In a panel approach of 19 OECD countries for the period 1965–1991, he finds that a proxy for openness has a significantly negative effect on capital taxes and a significantly positive effect on labour taxes. However, his results are not robust when one adds a qualitative dummy variable for international exchange rate restrictions and an interaction term of this dummy as a proxy for openness. A change in capital taxation is one of the main features in the more general discussion on the consequences of globalisation for the nation state. In their broad survey, Schulze and Ursprung (1999) consider the different links between the reduction in international arbitrage costs and fiscal policy. Their contribution covers topics such as the size of the public sector, the structure of public expenditures, the structure of taxes and the scope of redistribution policies. The survey comments on the reviewed literature as follows: “…many of these studies find no negative relationship between globalisation and the nation's ability to conduct independent fiscal policies.”1 The authors also suggest that the different hypotheses on the effects of globalisation are not mutually exclusive. They argue that some effects apply especially to tax policy, which should be distinguished from the effects on government expenditures. The plan of this paper is to focus on the following issues. First, we show that the proxy variables for corporate taxation used by Garrett (1995) and Quinn (1997) have conceptional difficulties and are responsible for certain counterintuitive results. Consequently, we use effective average tax rates on corporate capital with the methodology proposed by Mendoza et al. (1994). In this way, tax policy is better depicted, as we additionally control for tax base effects. Furthermore, we distinguish precisely between the measure of globalisation and country size. Second, we explicitly take into account the result of the tax competition literature concerning labour taxation and the impact of country size on the tax policy. To show the theoretical foundation, we rely on a small theoretical model of international tax competition. Third, since tax policy depends on a number of additional elements, we enlarge the simple tax competition model with the issues of dynamics, political preferences and uncertainty. We also include several sensitivity analyses to test the robustness of the estimated coefficients. We control for different macroeconomic variables, which should have an impact according to theory. Fourth, we use panel data for 14 OECD countries with annual frequency, covering the period from 1967 to 1996. To do so, we have collected and constructed data for Austria, Belgium, Canada, France, Germany, Greece, Italy, Japan, the Netherlands, Norway, Sweden, Switzerland, the United Kingdom, and the United States. Fifth, we expand our analysis of capital taxation on labour taxation and social expenditures. The objective of this paper is to explain tax policy behaviour. Before we start with the econometric analyses, we describe the variables that are important from a theoretical view. The main issue is capital taxation. As taxation of labour and social expenditures is linked to this topic through the government's budget constraint, further empirical results for these issues are added. Our empirical results confirm the theoretical prediction of a negative effect of international capital mobility on corporate taxation. Furthermore, as theoretical models of tax competition predict, labour taxation relative to capital taxation is positively related to increasing globalisation. We also find that government social expenditure responds positively to increasing globalisation. Thus, there is an evidence of different effects of globalisation at the same time. On the other hand, we observe a change in the tax-mix from the rather elastic corporate tax base to the rather inelastic labour tax base, which is a result of tax competition. On the other hand, we find increasing social expenditures, which seem to reflect governments' aims of compensating for increasing risk. In Section 2, we present the basic theoretical tax competition model, enlarged with further elements important for tax policy. Section 3 describes the data. In particular, the measurement of globalisation and capital taxation is depicted in detail. Section 4 regards the empirical results. Section 5 concludes.

Past empirical studies have found a positive relation between globalisation and capital taxes. Such results contradict the theory of international tax competition. In this paper, we provide results that clarify the past paradoxical outcomes. Using a panel of 14 OECD countries in the period from 1967 to 1996, we have reexamined the basic hypotheses of the tax competition literature. We have gone beyond the restrictive assumptions of the basic theoretical model by adding variables that account for dynamic effects, uncertainty and political preferences. Since we use effective corporate tax rates as the measure for capital taxation, our estimates apply to only a subcategory of capital tax competition. Our procedure is justified, since corporate tax bases are much closer to the assumption of the theoretical model than are other forms of capital taxes. We find that national governments lower corporate taxes as a consequence of increased globalisation, which, in contrast to past empirical studies, is consistent with the predictions of tax competition theory. The differences between our results and the results of past studies are due to our more appropriate measures for corporate taxation and openness, which consider tax base effects and avoid the bias of economy size. Second, as predicted by theoretical models of tax competition for a given level of public spending, we find a significantly positive relation between globalisation and the relation between labour and corporate taxation. Third, we find that governments' social expenditures are positively related to increasing globalisation. The change in the tax-mix from the rather elastic corporate tax base to the comparatively inelastic labour tax base supports the efficiency hypothesis of globalisation. On the other hand, the increase in social expenditures supports the compensation hypothesis of globalisation. The efficiency and the compensation hypotheses therefore both have a role in explaining government behaviour, the former for revenue, the latter for expenditure.