The Globalisation Debate Revisited : An Assessment of the « Constraints » of Fiscal Policies of the Nation-State

What does 'globaíisation' really mean? Ifwe were to evalúate ¡he concept of globalisafion in terms of the frequeucy oflts use iré would conchide that it refers to a concrete and unqiiestionable reality. tfowever, thls conclusión is farfrom adequate. Indeed, globaíisation is a very controversia! concept that has originaíed diverse political and academic debates. Reíd et ai (1999,) for exampie, in establishing a now classic trilátera! typology, have distinguished behveen aiithors it'/ío assuine an optimistic, scepñcal or critical visión in order to understand and explain thepotentia! impací of globaíisation on the Nation-State', as iré// as other dimensions. Startingfrom rhis t\polog)> as a framework for analysis, litis paper re-examines the specific debate on the impact of globalisation upon one ofthe more traditionally exclusive spheres ofthe Nation-State: fiscalpoiicies. !t concludes that ifreguiated according to 'econoinic invariable ¡aws', farfrom being an inevitably constrainingphenomenon, globaíisation is better understood ifit is observed as an econoinic poíiticalprocess (¡üstoñcaily determined), as insistently emphasized by sic political economy.


REVISTA DE ESTUDIOS INTERNACIONALES
The Globalisation Debate Revisited: An Assessment of the «Constraints» of Fiscal Policies of the Nation-State

Ricardo Camargo Brito
What does 'globaíisation' really mean?Ifwe were to evalúate ¡he concept of globalisafion in terms of the frequeucy oflts use iré would conchide that it refers to a concrete and unqiiestionable reality.tfowever, thls conclusión is farfrom adequate.Indeed, globaíisation is a very controversia!concept that has originaíed diverse political and academic debates.Reíd et ai (1999,) for exampie, in establishing a now classic trilátera!typology, have distinguished behveen aiithors it'/ío assuine an optimistic, scepñcal or critical visión in order to understand and explain thepotentia!impací of globaíisation on the Nation-State', as iré// as other dimensions.Startingfrom rhis t\polog)> as a framework for analysis, litis paper re-examines the specific debate on the impact of globalisation upon one ofthe more traditionally exclusive spheres ofthe Nation-State: fiscalpoiicies.!t concludes that ifreguiated according to 'econoinic invariable ¡aws', farfrom being an inevitably constrainingphenomenon, globaíisation is better understood ifit is observed as an econoinic poíiticalprocess (¡üstoñcaily determined), as insistently emphasized by sic political economy.
ledgement, there is substantial disagreement conceming the conceptua-lisation, causal dynamics, socio-economic consequences, implications for statepower and governance, and historical trajectory of globalisation.
One of the first comprehensive characterísations of such a debate was offered by Held et al, (1999) who propose a typology that distinguish among three main theoretical streams: 'hyperglobalisers', 'sceptics' and 'transformationalists'2. Economic globalisation has originated new forms of social organisation.
'Hyperglobalisers' include authors who characterize 'globalisation' as quite a novel phenomenon that would define a new era of human history in which the logic and discipline of the global market would not only be predominant but would also replace the historical role of the nation-state.Furthermore, in this sense, globalisation -as asserted by a popular commentator-would mean that «traditional nation-states have become unnatural, even impossible business units in a global economy» [Ohmae (1995): 5].According to most hyperglobalisers, the main cause of globalisation is the emergence of the global economy through the establishment of a sort of transnational network of pro-duction, trade and finance [Held et al (1999): 3], Furthermore, for those scholars, economic globalisation has originated new forms of social organisation that tend to replace the societal, economic and political role of the nation-state.The result is a view that assumes that the role of state is reduced to little more than a «transmíssion belt for global capital» or a «simple intermedíate institution sandwiched be-t\veen increasingly powerful local, regional and global mechanisms of governance» [Held et al (1999): 3].However, although most of 'hyperglobalisers3 share the premise that the market has imposed its rule over the state3, they can be classified into neo-liberals -those who assume the emergence of a single global market as an advanced stage of human progress [Ohmae (1995)], and neo-Marxists -those who highlight economic globalisation as an imposition of an oppressive global capitalista, which in many respects might have given rise to a more uneven and savage international order [Greider, (1997)].
It is worthwhile noting, however, that both approaches agree that globalisation is producing a new división of labour, explained by an increasing process of transnationalisation of production, that may have replaced the traditional North-South división.However, while the neo-liberal perspective assumes that global economic competition does not necessarily produce Other useful compilations on the globalisation debate are Scholíe (2005) andHolton (2005).In turn, for the best critique on the theoretical confusions surrounding such debate, see Rosenberg (2000).«...the impersonal forces of world markets, integrated over the post war period more by prívate enterpríse in finance, industry, and trade than by cooperativo decisions of governments, are now more powerful than the states to whorn ultímate political authority over society and economy is supposed to belong» (Strange, 1996, p. 4).zero-sum outcomes (because countries are able to specialise in accordance with their comparative advantages, and scale economies, generating enough wealth that would allow winners to compénsate losers and still remainbetter off), neo-Marxists stress that globalisation has not only failed to resolve global inequalities, but has also created and reinforced structural patterns of inequality between and within countries.
The most predominant feature of the current international economy is the process of economic regionalism.
'Sceptics', inturn, comprise scholars who definitely refute the idea that globalisation has given rise to a process of integrated worldwide economy in which states would be increasingly redundant.Instead, sceptical commentators point out that globalisation is, at best, a process of internationalisation, which according to thern means no more than a process of interactions which take place predominantly between nation-states [Hirst and Thompson (1999): Ch. 4].Indeed, for sceptics, globalisation is fundamentally a myth in two main senses.First of all, current levéis of global economic integration would not really be historically unprecedented.On the contrary, they would be less intense than world flows of trade, investment and labour in the nineteenth cent ury.Second, contemporary globalisation, far from creating a completely integrated global market, would be focused solely on the 'triad' formed by the three main eco-nomic blocs of the European Union, NAFTA and APEC [Ruigrok and Tulder (1995): 148-51;BoyerandDrache(1996); Hirts and Thompson (1999): Ch. 4].This process of economic regionalism, which is purportedly contradictory to the globalisation process, is really the most predominant feature of the current international economy [Weiss (1998)].
Moreover, hyperglobalisers's insistence in defining globalisation as a process that brings about a new, less statecentric order [Held et al (1999): 5] is strongly rejected by the sceptic's thesis, that not only denies that the nation-state is in retreat, but also considers that states are the real promoters of internationalisation.It is alleged that this is particularly true in the case of the United States, a superpower that after the Second World War began supporting a multilateral economic order, abandoning its formerly traditional unilateralism This is considered to be the real cause of the beginning of an increasing process of liberalisation of national economies worldwide, and not the other way around [Gilpin (1987)].Naturalíy, sceptics admit that internationalisation might constrain the range of action of governments, but for them, such constraints do not immobilize governments at all.In fací, in many cases internationalisation of capital might «not merely restrict policy choices but expand them as well» [Weiss (1998): 184].
For sceptics, a far more important fací are the consequences that a process of economic internationalisation would produce in an international order, such as a huge increase in the economic margina-lisation of many developing countries as a direct result of extreme concentration of trade and investment flows within developed countries.Furthermore, some sceptical commentators also refute the thesis of the emergence of a new interaational división of labour, linked to theNorth's deindustrialisation and the extensión of new webs of production in the South.Indeed, these authors emphasize that the belief that the world economy is dominated by global corporations without regard for any traditional jurisdiction is another «myth».In fací, the patterns of foreign investment flows show clearly that they are mostly concentrated in developed countries and that most multinational corporations are, in fact, strongly linked to their home state orregions.Therefore, farfrom considering that a new global economic order has emerged as a consequence of internationalization, sceptics assert that in structuraí terms, the deeply rooted patterns of inequality and hierarchy of the world economy have only experienced marginal modifícations overthe last century [Held et al (1999): 6].Despite hyperglobalisers belief, sceptics argüe that increasing patterns of inequality in the contemporary world economy are failing to genérate a global civilization.On the contrary, they have contributed to the rise of both fundamentalism and extreme nationalism, which, in turn, have led to a world fragmented in cultural and ethnic enclaves.
Finally, 'transformationalists' are a group of scholars who agree with hyperglobalisers in assuming that globalisation is a historically unprecedeníed process experienced by societies and gov-ernments around the world, characterísed by a weakening of the distinction between international and domestic affairs [Rosenau (1990); Cammilleri and Falk (1992); Ruggie (1993); Linkiater and MacMillan (1995); Sassen (1996)].
The patterns of inequality and hierarchy of the world economy have changed marginally over the last century.
Furthermore, they view globalisation as a central driving forcé that is modifying the social, political and economic aspects of modern societies and the world order [Giddens (1990); Scholte (1993Scholte ( ) (2005)); Castells (1996)].Indeed, Giddens (1996) for instance, characterises the changes generated by globalisation as a 'massive shake-out' of societies, economies, institutions of governance and the whole world order.However, while hyperglobalists assume that the final product of globalisation will be the emergence of a single global market and the disappearance of the nation-state, 'transformationalists' argüe that the direction of globalisation is both uncertain andunpredictable, preventing a reliable forecasting [Held etal (1999): 7].Indeed, globalisation is characterísed as a contingent and openended historical process that has countless variables, most of which would have contradictor/ effects.
It is also worthwhile noting that 'transformationalists' do not consider that the emergence of a single global system is equivalent to a balanced and equal world society.By contrast, they associate globalisation with the emergence of new patterns of global stratifícation, which, in turn, determine different levéis of social, community and national integration of the global order.The rise of a new form of global power relations, which is a central point of the transformationaíist thesis, might have replaced traditional characterisations such as the North-South división or the First and Third worlds that, according to transformationalists, would have become obsolete.In fact, these oíd characterisations of the international scenario oñen overlook the way in which globalisation has modified patterns of inclusión and exclusión between countries.These might now be better described as networks that cut across and penétrate all societies and regions of the world [Hoogvelt (2001): xü-xiv].
A new form of global power relations might have replaced characterizations such as the North-South división.
ForCastells (1996) andRuggie (1996) the main cause of this new stratified global order is the de-terrítorialisation ofeconomical activities, that is a direct consequence of an increasing process of transnationalisation of production and finance, and as a result of which the national economic space no longer coincides with national territorial borders [Castells (1996): Ch. 2:77-147].
Compared to sceptical and hyperglobalist approaches, 'transformationalists' assume neither the obsolescence ñor the unchanging condition of the nation-state, but a new phenomenon called the juxtapositionofauthorities [Helde/L<:¿/(1999): 8].Indeed, although 'transforma-tionalists' still argüe that states retain the ultímate legal claim to exercise effective supremacy over what occurs within their own territories, they also emphatically argüe that states are being affected by the expanding jurisdiction of international institutions and the constraints imposed by international law [Goodman (1997)].Inthis sense, the traditional conception of sovereignty as an undisputed authority within a given territory would be outdated as a descriptive statement, and would have been repíaced by the notion of a juxtaposed state, that is, a state wherein different kinds of transnationaljurisdictions overlap [Heldeí al (1999): 8].In other words, the state would not disappear but the relationship between sovereignty and territoriality might be dramatically transformed.

GLOBALISATION AND THE NATION-STATE
As we have seen, one of the most controversial issues within globalisation literature relates to the state.States have been the most important actors in the political organisation of societies and have played a key role in the external relations between nations during the modera era.For this reason, the way in which globalisation affects them, either by producing the obsolescence of their functions, the reconfiguration of their power, or even the continuiíy of theirtraditional sovereign status, has been object of a huge discrepancy among scholars in the globalisation debate.
A real global system of independent states started to emerge as a direct consequence of the decolonization process.
Modera states emerged in Western Europe and its colonial territories in the eighteenth and nineteenth centuries, but their real origins date back to the late sixteenth century4.The characteristic that distinguíshed the first states from earlier forms of political organisation was the symmetry and correspondence existing between sovereignty, territory and legitimacy.Indeed, the development of the concept of sovereignty as a special claim to the rightful and exclusive exercise of political power over a given territory was a key factor in the emergence of the state [Held et al (1999): 9].As it is known in the modern era, the natíon-state has developed a unified system of rule, a central administration, a concentrated mechanism of fiscal policies, a system of lawmaking and law enforcement, a professional army, as well as a diplomatic corps in charge of its foreisn relations.The consolidation of the nation-state fírstly took place in Europe, but it was not until the aftermath of the Second War World that a real global system of independent states started to emerge as a direct consequence of the decolonisation process.Furthermore, the final consolidation of this global system of states only happened during the late twentieth century, when the great empires -European and Soviet-finallycollapsed, Therefore, from a long-term historical perspective it seems undisputable that the modera nation-state appeared at the end of the twentíeth century as the predominant type of political organisation of nations worldwide.Inthis sense, sceptics claim that the twentieth century is the age of the modera nation-state and, consequently, radically reject the possibility that the state is in retreat, would appear to be backed by solid evidence.Indeed, it seems undisputed that modera nation-states have concentrated the legitímate exercise of the use of forcé andjudicial regulation, established permanent military forces, implemented central fiscal policies, and imposed an official language and a national identity.Even some hyperglobalists recognise this point5.
Furthermore, many states, especially inthe Organisation forEconomic Co-operation and Development (OECD), have traditionally ímplemented macroeconomic In fact, theír origins are often referred to as the Peace Treaty of Westphalia in 1648, which concluded the Thirty Years' War.See Falk (1969) and Kehoane (1995).«The intrusión of governments into our daily lives in the 1990s, as compared, say, with the 1890s, is palpably greater.Statutory or administrad ve law now rules on the hours of work, the conditions of safety in the work place and in the home, on the behaviour of citizens on the road.School and uníversities are subject to more and more decisions taken in ministries of education.Planning offícials have to be consulted before the smallest building is started or a tree is cut down», in Strange (1996), p. xi.strategies, shiñing from Keynesian demand management in the 1950s and 1970s to extensive supply-side measures in the 1980s and 1990s, in order to promote economic growth and widespread employment [Held and McGrew (2000): 10].In spite of the fact that most of these strategies have often been unsuccessful, they seem to demónstrate the central role played by the state in controlling the national and International economy.
Whether or not international finan cial and commercial constraints preclude governments from setting up national economic strategies contínues to be debated.Furthermore, it is worthwhile noting that 'sceptics' have always recognized that the role of the state has historically been constrained by the prevailing structure of the world order, that is, the modern nation-state system and capitalist economic relations.However, they dispute the 'hyperglobalistV assumption that the key characteristics of globalisation are the increasing restrictions imposed by the dynaraic of the international economy over a set of policy options confronted by governments.It has been argued that a good example against this type of general constraints imposed by globalisation might be the common shape of policies implemented by OECD countries as well as by states in sub-Sabaran África, east Asia and Latín America in the late twentieth century such as market liberalisation, welfare cut-backs, minimal regulations of prívate capital flows and processes of deregulation of labour markets, wbich far frombeing attributed by sceptics to just one inescapable single global driving forcé, result from a diverse set of political factors [Held and McGrew, (2000): 10].This point, however, is far from uncontested.Moreover, to what extent the implementation of these policies has been a consequence of either the constraints produced by globalisation or a result of an autonomous decisión of the government or any combination of both, is a central unresolved matter in the literature concerning globalisation and the nation-state6.In fact, while international fmancial mobility, especially capital mobility, might impose similar economic restrictions on all national economies around the world, governments respond to this constraint in different ways.The various responses would be explained, amongst other factors, by variations in the type of capitalism predominant in the national economy of a given state, by the type of political system of the countries and by the level of social integration of national societies.The current stage of the discussion, however, continúes to be whether or not international fínancial and commercial constraints preclude governments' ability to set up national economic strategies.In other words, is it true that as a consequence of international capital mobility, govern-For a critical review of this debate see Saad-Filho and Johnston (2005).See also Palan and Abbot (1999), Desai (2001) and Bruff (2004).
ments for instance, have been forcea to adopt similarneoliberal economics policies, which, in turn, have consequently stimulated fínancial discipline and reduced the margin of governments' actions?Moreover, has the global competition in which states are involvedto attract international capital to national economies, increasingly reduced governments' ability to maintain high levéis of social protection or welfare state programmes?In short, are we in the presence of the definite end of the welfare state as (we have known' it, as former President of the United States, Bill Clinton has declared?It is uponthis debate about the ability of the state to pursue national economic strategies and to maintain a welfare state systern free of the constraints imposed by the global economy that this paper will be focused, and specifically on the impact of globalisation on the fiscal policies of the nation-states.

GLOBALTSATION AND FISCAL POLICIES
The main discussion of the literature on globalisation and fiscal policies focuses on the claim that a process of capital-market integrationwould produce aphenomenonof 'fiscal squeeze' [Grunberg (1998): 591].A fiscal squeezewouldhavetwo componente: fírstly, the increasing difficulty of governments to collect revenue from the prívate sector, and secondly, a rise in demand for public spending.These elements constitute the hypothesis of «double jeopardising» of public finance [Grunberg (1998): 597].

Market dislocations and income
inequalities might stimulate the demand for social insurance from more vulnerable social groups.
Grunberg identifies three áreas of economic globalisation-trade, investment and financia!liberalisation-aíí of which have direct and indirect effects on public bud-gets7.The effects of economic globali-sation8 are summarised as the increasing Firstly, trade liberalisation leads directfy to a loss of revenue from tariffs, the reduction or elimination of which is the main target of a trade liberalisation policy.The indirect effect of trade liberalisation on the public budget is the demand and need for additional public resources to either subsidise fírms or to provide social adjustment, Secondly, fínancial liberalisation brings as a direct effect the loss of revenue from 'fínancial repressíon'.Indirect effects include the loss of seignorage revenue as a consequence of the lower inflation brought by fínancial liberalisation, the revenue íost by governments from a disappearing so-called «inflation tax» (that is, the ability to repay investors in devalued currency), and the increase of governmental spending motivated by ñnancial crises, which in the last few years have become recurrent in developing countríes.Finally, investment liberalisation produces as a direct effect the challenge to tax collectíon.The best illustration of this concern is foreign direct investment taxation.The indirect effect of investment liberaüsation is tax competition.Since capital tends to prefer low-tax environments, states are engaging in a competition to lower corporate and capital gains taxes, resulting in overall shortfalls in tax receipts, Grunberg (1998), pp.594-595.Economic globalisation includes trade, investment and fínancial liberalisation.
The Globalisation Débale Rerisiled difficulties in raising reverme, added to the growing demand for public spending [Grunberg (1998): 601].In turn, ithas been argued that the main political consequence of economic globalisation is the breaking of the social deal that tacitly operated between govemments and the working class duríng the post-Second War World period, on the basis of which govemments committed themselves to establishing social insurance in exchange for foreign trade liberalisation [Rodrik (1997a): 2].In this sense, Rodrik (1997a) warns us about social disintegration and the loss of domestic consensus in favour of open markets that might result from an unbalanced process of economic integration.
The 'doublejeopardy» on fiscal policies produced by globalisation has also been analysed on the basis of the so-called efficiency and compensation effects9.The formerrefers to the impact of globalisation on the suppiy side of the political market.In other words, it focuses on the government's ability to fínance public goods in a globalised era by íncreasing taxes.The latter, in turn, assesses the influence exerted by globalisation on the demand side of the political market.In this sense, the demand for public spending, especially for income transfers, is allegedly positively correlated to a hígher degree of economic globalisation.This is because market dislocations and the counteractionof income inequalities generated by globalisation might stimulate the demand for social insurance from the more vulnerable social groups.
The point under discussion is whether or not the compensation effect, which would increase public spending, is balanced and additionally, constrained by the efficiency effect.Although this discussion is still in a preliminary stage, most commentators agree that a final assessment of the efficiency and compensation effects of globalisation is an empirical matter [SchulzeandUrsprung, 1999, p. 301).In the next section, I will turn to the theoretical discussion as well as to the more relevant ernpirical results concerning globalisation in both the suppiy and demand side of the political market.
Govemments impose taxes on capital so as to maximize social welfare.

GLOBALISATION AND TAX COMPETITION
The basic contention of globalisation and tax competition literature is that in a globalised era, while many jurisdictions compete for a given stock of capital, which is assumed to be perfectly mobile across jurisdictions, and pursuing the highest net of tax return, labour is relatively irnmobile.
Consequently, a theoretical perspective asserts that if all jurisdictions have equivalent technology and not considering scale economies, factors' gross-of-tax returns depend only on their relative scarcity [Schulze and Ursprung (1999): 309].Moreover, in the case of capital taxation, if no taxes are levied or different jurisdictions are subject to the same rate of taxation, capital wouldbe allocated effíciently and governments would receive identical gross rerura in each jurisdiction.It would therefore be expected that governments impose taxes on capital so that social welfare is maximised.In other words, governments would seek to créate the best conditions for improving the profíts of a representative agent10.
However, since there is tax competition, the main conclusión in the relevant literarure is that given capital mobility, public goods will be underprovided and therefore an ineffíciency condition will be generated, which would only be resolved by a process of worldwide tax harmonization".There are at least two main factors that play against tax harmonisation.The first is the political reality of an international system, composed by sovereign nation-states that have usually opted for autonomous national policies instead of internationally coordinated taxation.The second factor concerns the more favourable position of smaller countries vis-á-vis bigger countries.In fact, the former countries benefít from an extemality: a higher foreign tax rates, and a bigger the tax base of small countries, reduces the net return to capital [Schulze and Ursprung (1999): 310].On the contrary, a reduction of the tax rate in large countries would result in a higher world net return to capital but not in the attraction of capital from small countries12.In this sense, if country size differences are too large, small countries will resist tax harmonization.
If country size differences are too largCj small countries mil resist tax harmonization, Therefore, it has been argued that in a scenario of capital mobility and disequilibria of international tax policies -assuming also that the residence principies would not be enforceable13, and other tax instruments are available -an effíciency per-This utility, in turn, derives from two main factors.Firstly, from public goods provided by the state and fínanced by tax revenues, and secondly from prívate goods, which are purchased out of net income.The effíciency is reached when the marginal utility generatedby public goods consumption equals marginal utility of prívate goods consumption.This equation is called the 'Sarnuelson condition1, In fact, if it is assumed that governments behave rationally, marginal social benefit of taxation (MSBT) isequal to its marginal social cost (MSCT).Eitherin a scenario ofimmobile capital orperfect capital mobility, MSBT is always the same, that is, the benefít generated by the provisión of additional public goods.However, MSCT varíes in a situation of mobile or immobile capital.Indeed, in the latter case, MSCT equals the reduction in prívate (after-tax) income, however, in the former case, MSCT additionally includes the loss of capital, which leaves the couníry as a consequence with a higher tax rate on capital.In other words, MSCT in one country produces a positive externality for other countries because ií increases their tax base and incomes.Tliese ineffíciency conditions might be eliminated by an international coordination of capital taxrates, see Schulze and Ursprung (1999), p. 309.The reason of this disequilibria lies in the fact that a large country is able to levy at a higher tax rate because the consequent negative effect of the tax base (by capital exit) is smaller in per capita terms and therefore, a higher tax rate translates into higher revenue per unit of capital, but a smaller tax base.The underlying reasoning ¡s that taxes should noí necessaríly distort production effíciency, if taxes spective indicatesthatmobilecapitalshould 4.1, Empirical evidencefor be altogether exempted from taxation and globalisation and taxpolicies immobile domestic factors (land, labour) should be taxed instead [Schulze and xhe general prediction derived from Ursprung (1999): 310]14.In this sense, it literatee is that capital taxation will be would be expected that capital tax rates negatively related to the degree of Ínterfall with the increase of capital mobility.national integration generally, and to inter-However, taxes are set in a political pro-national capital mobility in particular16.This cess, largely influenced by distributional means that the higher the degree of interconsiderations, which makes it difficult to national integration and capital mobility, the what extend this trend is due to capital iower the capital tax rates.mobility orto ideologicalconsiderations15.
m factj Rodrik (1997) examines the influence of a country's openness on labour and capital income tax rates sepa-rately17.Rodrik's main fíndings are that on activiíies and commodities can be set optimally, that is, taxes do not interfere with factor or commodity trade.This, in turn, implies that capital income should be taxed according to the residence principie, see Diamond and Mirrlees (1971) for a discussion, 14 An efficiency perspective disregards distributional effects, which are the main cause of concern of the «doublejeopardy' hypothesis [Grunberg (199S) and Rodrik (1997)], and also that governments behave efficiently, that is, providing the optimal level of public goods as implied by the Samuelson rule.13 Furthermore, Richardson (2002) has outlined two theoretical reasons by which capital taxes can increase in a scenario of capital mobility.Firstly, since in an open econorny the double taxation problem is less important than in a closed economy, we wouid expect higher taxes on capital income.In a closed economy capital income is íaxed as it accrues to corporations and as it is dispersed to a company's owners, both of which are national agents.For this reason, no-distortionary taxation usually includes low corporate taxes.In the case of an open economy, there are foreign investors, which means íhat national governments are less concerned with a double taxation and mus capital tax rates tend to be higher.See Richardson (2002), p. 14.Secondly, the location decisión of FDI around the world is not only influenced by wages and capital tax rate, but also by a set menú of services, which includes such factors as skilled labour, reliabie infrastructure and political security.In this sense, Ríchardson (2002) concludes that international competition over footloose fírms might be manifest not in tax cutting but in the provisión of the best set of productive services.It would be a sort of market competition for FDI, in which different jurisdictions offer di verse menus of taxes and services, among which corporations choose their location according to their preferences over those menus, Huíbauer (1999) Cf. in Richardson (2002), p. 16. 16 In this sense, Schulze and Ursprung (1999), have proposed three hypotheses: a) «At a given point of time, the countríes which tax capital least are the ones whose econormes are most integrated in the world capital markeí.b) Capital taxation has been reduced overall over the last few decades as countries have becorne increasingly integrated.c) Similar couníríes have similar tax rates whereas smaller countries have lower rates than bigger ones».See Schulze and Ursprung (1999), p. 312 17 Opetvuess is defíned as the sum of exports and ímports divided by GDP, using data of IS OECD countries, between the years 1965 and 1991 (Australia, Austria, Belgium, Canadá, Denmark, Finland, France, Germany, Italy, Japan, Netherlands, New Zealand, Norway, Spain, Sweden, Switzerland, the rate of openness is positively and significantly correlated wíth labour tax rates; and is negatively and signifícantly correlated with capital tax rates.He concludes that, «there is a strong evidence that as economic integration advances, the tax burden... is shifted from capital to labour» [Rodrik (1997): 20].However, these fíndings are limited by the fací that the openness variable refers to the goods trade rather than to capital mobility, that most literature links with tax competition18.Notwithstanding, Rodrik (1997) uses an additional dummy variable to capture the restrictions on capital mobility as well as to analyse the interaction of this dummy with openness.Results show, in this case, that the direct effect of openness is still signifícantly positively correlated with labour tax rates, but this effect is less strong in the presence of capital account restriction.However, the direct effect of capital account restrictions on labour tax rates is negative, but negligible.Furthermore, the direct effect of openness is still negatively correlated with capital tax rates, but no longer significant.In turn, the interactive effect of openness and capital account restrictions -both variable acting together-on capital tax rates is signifícantly positive and larger than the abovementioned direct negative effect.The direct effect of capital account restrictions on capital tax rates is, on the contrary, negative.This dummy variable is taken from the summary table in the IMF's annual reports on exchange arrangements and exchange restrictions [Rodrik (1997): 18, note 5].
In short, Rodrik's empirical findíngs on the relationship between trade liberalisation-capital mobility and the taxation structure show, though not conclusively, that an increased tax competition shifts the tax burden away from mobile towards iramobile factors such as labour.
The direct effect of the capital mobility ratio does not show a significant correlation with capital taxation.Garret (1995), in tura, examines the relationship between capital taxes as a share of GDP and lagged endogenous variables such as growth, unemployment, the share of trade, and an Índex of capital mobility19.He fmds that only an increas-U.K. and U.S.).In íurn, as a tax rate measure is used the 'effective average tax rates' as calculated by Mendoza et al (\ See, Rodrik (1997), pp. 19-20.It has been argued that the trade share is referring to the actual flow of goods whereas capital mobility refers to a potential to move the production factor capital rather than the actual magnitude of flows.Moreover, the trade share is influenced by the country size independent of the degree of capital mobility and thus openness is a crude and biased proxy for the degree of capital mobility.See, Schulze and Ursprung (1999), p. 314.This Índex is elaborated by classifying the number of types of international capital account transactions and by assigníng a given rank from 1 to a máximum of 4.However, this índex does not measure the actual strength of restrictions, but proxies only the coverage of restrictions, Garrett stresses that it is roughly in accordance with the saving investment coefficient.However, other studies dispute that this coefficient is a good indicator for the degree of capital mobility, see Jansen and Schulze (1996).
The Globalísation Debate Revisíted ing exposure to trade is positively corre-associated with financial liberalisation in latedwith capital taxation.In other words, the majority of regressions.In particular, the direct effect of the Índex of capital corporate taxation reverme as a percentraobility does not show a significant cor-age of individual taxation is significantly relation with capital taxation.On the other positively correlated with trade balance as hand, it is worthwhile noting that the in-a percentage of GDP and financial teraction of trade-share with an added in-liberalisation, which indicates a re-distribudex that captures the influenceof the Left tive aspect of taxation.Quinn also fmds as a political variable is positively corre-that corporate taxation revenue as a perlated with an increase in capital taxes, centageof GDP forboth OECD andnonshowing an interesting relevance of politi-OECD countríes is positively and significal factors in favour of the compensation cantly associated with financial hypothesis.
liberalisation [Quinn (1997): 539, Table 3].Quinn (1997) studies the effects of Swank (1997) has studied the impact variations in international financial regu-of domestic and international explanatory lations20 upon corporate taxation21.The variables on corporate profít taxation and results of his research show that most ernployer social security and payrolltaxacorporate taxation revenues are positively tion22.His main findings are that all of the The data used is 15 OECD countries during the years 1967 and 1990 (Austria, Sweden, Finland, Norway, Denmark, Germany, Belgium, Italy, Netherlands, United Kingdom, Australia, Canadá, Japan, United States and France).Additionally, Garrett considera an Índex for the partisan centre of gravity in cabineí and the legislature, which is also analysed in íts interactíon with íhe índex of capital mobility and the trade share, respectively.See Garrett (1995), p. 658. 20 The change in international financial regulations is a coding rule, which, following the categories of The First Annual Report on Exchange Restrictions of the IMF, contemplates restrictions on both currency transactions and the underlyíng international commercial transactions.Inward and outward capital account transactions are measure by Capital, which is scored on a 0-4 scale.real GDP, (iv) the consumer price Índex, (v) an índex for the political power of the left, (vi) a dummy for election years, (vii) total government outlays, and country dummíes.The ¡ntemational variable ís capital mobility which is captured by three different measures: a) the actual total capital inflow and three measures that he uses to capture capital mobility are systematically, signíficantly, andpositively correlated with corporate taxation, which is coincident with results of Quinn's research23.However, contrary to results reached by Quinn (1997) and Garrett (1995), he fmds that openness, captured by the share of trade ín GDP, is negatively correlated with capital taxation.

An assessment of the empirical evidencefor globalisation and taxpolicies
To explain the sometimes 'intersecting' results of the abovementioned studies, it is worthwhile examining the various indicators used to set up a specifíc definition of globalisation as well as of taxpolicies.
Capital Mobility.Even though the measure of globalisation used by Garret (1995) and enhancedby Quinn (1997) and Swank (1997) has been accepted as the best way to capture the degree of capital mobility in the contemporary scenario of economic integration24, it still retains one importan!conceptual shortcoming: this measure disregards the importance of transferpricingpossibilities.Indeed, ithas been argued that even if capital flows were excessively contrallad and restricted, and arbitrage conditions did not exist, profíts could still be shiftedbetween subsidiaries [Schulze andUrsprung (1999): 313, note 26] and Quinn's indicators of capital mobility would be unable to capture such variations.This would become a more significant weakness in the near future, considering that a central feature of the current stage of capitalism seems to be a significant development of the intra-trade fírm [Castells (1996): 172].
The alleged correlation between capital mobility and capital taxation should be carefully analyzed.
Taxation Competition.Garret (1995), Quinn (1997) and Swank (1997) use only one measure to capture the degree of tax competition: the capital income taxation revenues25.This measure also presents a fundamental conceptual problem: theposioutflow as a share of GDP, b) an índex of restríctions of capital account transactions (Score 0-4), c) a broader índex of capital account as well as exchange restríctions (score 0-14).Both Índices were taken from Quinn (1997).The data used is 17 industriaüsed countries duríng the years 1966 and 1993.23 They are also, even though to in a lesser extent, positively correlated with employer social security and payroll taxation.24 Edwards (2000) compares Quinn's indicator with an Índex based on the number of years within a certain period that, according to the IMF, a particular country has not ímposed capital controls.This is called The Number-of-years-with-conírols: NUYCO is an índex used by Alesina, Grílli and Milesi-Ferreti (1994) and by Rodrik (1998).Edwards fmds two main advantages of Quinn's indicator over NUYCO: it allows for the capturing intermedíate situations of a country's capital account openness and it is available for two different periods of time.See Edwards (2000), pp.9-10.25 Tax base multiplica by tax rate.
The Globalisation Débale Revisited tive correlation found between Corporate Income Tax revenues and capital mobility could have as an alternative explanation the increase of the tax base, as a consequence-for instance-of the riseof operating profíts as a share of GDP26, and in fact, would not correspond at all to an enhanced tax competition process27 [Schulze and Ursprung (1999): 316].Therefore, thepositive correlation asserted by Garret, Quinn and Swank that exists between capital mobility and capital taxation should be carefully analysed.
Ideológica! postures asserting that globalisation only affects tax policies in one >vay should be observed with caution.Rodrik (1997), in turn, uses the effective average tax rate taken from Mendoza etal (1994) as a measure of capital taxation, which seems to capture more properly the strategic variable, namely Corporate Incomes Tax rates and also has a more effective control on the tax base effect.In fací, the effective average tax rate presents four advantages over other measures of tax ratios: (i) it includes the actual amount of tax collected in the numerator, and thus implicitly takes into consideration the combined effects of statutory rates, tax deductions and tax credits; (ii) it is much simpler to calcúlate than marginal effective tax rates; (iii) it is a macroeconomic indicator and therefore is much easierto compare by country and is also an appropriate input for large-scale macroeconomic models, and (iv) it might provide a guide to tax reforms as they offer more insights than the distinction between direct and indirect taxes, or the statutory íax rates [Volkerink and Haan (2000): 11].The calculation of Mendoza et al (1994) has been criticised because it gives rise to the misleading condition of the use of the personal income ratio (the ratios of personal income tax revenues over the sum of wages and salaries, the operating surplus of prívate unincorporated enterprises and property, and entrepreneurial income) as an intermedíate step in calculating labour and capital income tax ratios.Indeed, incomes from labour, capital income and transfers are included in the 'tax base' of personal incomes and it is assumed that the same average tax rates apply to all these income categorías.The problem with this assumption is that in many OECD countries some income components are largely exempted from taxation or have different statutory tax rates.Trying to tackle this shortcoming, Volkerink and Haan (2000), modifying the methodology of Mendoza et al (1994) and recalculating the tax ratio, fmd that neither the trend ñor the level of the labour The 19SOs were very profítable years for corporations.Indeed, the operating profíts in 15 EU member states as a percentage of GDP increased from 9.7 per cents to 13.0 per cents between the years 1980 and 1996, cf.Kramer (199S).This is because revenues are not governments' strategic variables; on the contrary, they are consequences of the application of strategic variables such as the capital tax rate.Therefore, at best, they indirectly reflect the tax competition process in which governrnent could be involved.income tax ratio or the capital income tax ratio are well represented in at least some of the countriesanalysedbyRodrik (1997) [Volkerink and Haan (2000); 25 -47].
Apart from Volkerink and Haan's critique of the effective average tax rate measure used by Rodrik (1997) it is also worthwhile highlighting as we did before that Rodrik's fmdings are limited by the use of a commodity trade variable rather than capital mobility as a measure for globalisation.
The basic premise of apolitical models is that state intervention in economic spheres is only intended to correct marketfailures.
In sum, the empirical fíndings analysed genérate a far from conclusive result, mostly due to important theoretical and methodological shortcomings.Because of this, the debate on the fíeld must be observed with caution, avoiding those ideological postures that unreflectively assert that there exists only one irrefutable way in which globalisation affects tax pólices.That said, it is worth noting that even given the limitations of the empirical research analysed, the relevant literature seems to point out to the following main conclusions on the relationship between a given measure of globalisation and taxation: a) It might reveal that during the 1980s and 1990s in some of the countries subject to analysis capital tax revenue was stable and capital tax rates showed a modérate downward trend; b) Apparently the hypothesis of the race to the bottom has not taken place, since globalisation has not led to a reduction of capital tax revenues.c) However, effective tax rates on labour have risen substantially and at least the relative contribution of capital tax has been reduced.This means that the impact of globalisation on tax structure is in accordance with the effíciency effect, implying that the cost of mobile factors, notably labour, has increased with globalisation [Schulze and Ursprung (1999): 321].
However, a real assessment of the way in which the effíciency and compensation effects of globalisation have been manifest either in a competing or complementary way, also demands a review of the expenditure side of fiscal policies also.

GLOBALISATION AND PUBLIC SPENDING
Literature conceming globalisation and public spending is focused on two main theoretical aspects.The fírst refers to a more general relationship between globalisation and public spending, andis illustrated by questioning to what extent government spending is determined, and how government size is modified, by globalisation.The second point of concern focuses on a more specific relationship, and is pointed out by questioning how globalisation affects the patteras of government spending.These aspects will be analysed separately in the followings sectionofthispaper.

Government growth and globalisation
The impact of globalisation on government gro\vth has been studied on the bas is of two theoretical approaches: apolitical and political-economic models.

L1. Apolítica! models
According to the basic premise of apolitical models, state intervention in economic spheres is only intended to correct market failures.In this sense, government growth would only reñect a public view of the needs of a given society, based on a standard consumer theory28.Since literatura on the field has established that government spending depends on, or it is influenced by, input prices29, incomes30, preferences, population size31, community inputs32, and the increased cost of tax collection, it has been argued that if globalisation affected any of these factors, it would also indirectly alter government spending [Schulze and Ursprung (1999): 323].
Let us consider now some hypotheses applied to the abovementioned factors in order to discover whether or not it is possible to make a general prediction about the impact of globalisation on government spending according to this model.
First, it seems that globalisation, as expected, would tend to restrict the increase of input prices of public ser/ices due to the reduction of wage rates produced by price equalisation and migration.Also, competition for mobile factors cansed as aconsequence of globalisation might act as an incentive for the development of more capital-intensive technologies in providing public services.In turn, if we assume that globalisation originates an increase in income volatility, it would also stimulate an increased demand for public welfareprogrammes.In conclusión, all changes in the variable triggered by globalisation would thus be reflected in a positive impact on public spending33.
The standard consumer theory assumes that consumer utility is a funcíion of final service output.The final service output depends on publícly provided inputs and community inputs.Empirical studies confírm that input prices significantly influence public spending.The relationship is as follows: since input prices have risen considerably and steadily since the 1930s, and because public inputs represent mostly services (low capital-labour intensities) and techmcal progress has augmented capital, and if demand is price inelastic, then government growth can be explained by the increase of input prices.See Cf.Ferris and West (1996).Several empirícal researches find a relationship between national income and a larger number of government services.For a review of the most recent literature, see Richardson (2002).The relationship between population size and public spending has not clearly been established.However, some investigations -see Borcherding, T. (1985)-have concluded that the effect of population size on public spending is not really signifícant.In any case the effect is not easy to measure.Community inpuís are traditionally studied as substitutes for public inputs to final services outputs.Community inputs depend on socio-economic variables, basically the female labour-market participation rate and the household structure, and therefore any variation of these factors may have an important influence of public spending.See Schwab and Zampelli (1987).SeeKatsimi (1998) for a theoretical andempirical study of the relationship of income volatility and Therefore, if globalisation produced a decline of community inputs, one could reasonably assume an increase in public spending as a necessary palliative of the vacuum left within the community [Schulze andUrsprung (1999): 325].
Ambiguities start when \ve examine other factors, such as consumer preferences.Indeed, it is undisputed that changing preferences might affect the demand for any goods, including public spending.Therefore, although preferences are usually assumed to be invariable, if globalisation brings about changes in a country's preferences, these changes must be considered in the analysis of government growth.Either they increase or decrease it34, However, there is scant literature in this área and the effects of these changes are unknown.
Furthermore, the analysis of tax collection costs points to the opposite trend.Indeed, since predictably globalisation would increase the costs of tax collection, it would therefore result in a downward pressure of public spending.
Consequently, due to the limited power of predi ction of apolitical models, it would be too risky to assume a definite trend in the irnpact of globalisation on government spending unless explicit political factors are considered.
Acording to some authors the influence ofnational pressure groups will be reduced if globalisation produces changes in political decision-making.

5,1.2. Political-economic models
These models interpret public spending as a result of policies of social redistribution.The basíc assumption is that public goods produced by governments are equally distributed amongst voters and fínanced by proportional income taxes35.The mainprediction ofthis model is that if we assume that in a globalised scenario income inequalities were to rise36, ceterís paribiis, we would expect an expansiónary effect on the government share of GDP.However, using, the same models other authors have argued that if globalisation brings about a substantial change in the state's political decisiónsocial insurance government expenditure in the context of a closed econorny, and Rodrik (1998) for the analysis of the same relationship for open econornies.Changíng country's preferences could be due to migration or increased exposure ío other political cultures.See Schulze and Ursprung (1999), p. 325.The preferences of the electoraíe are decisive to determine the level of government spending.Within the elecíorate the median voter decisión seem to be the most decisive.As the median voter decisión varíes positively with the difference between mean and median incomes, government spending is fínally determined by this income difference.However, the ernpirical results ofthis theory have not ahvays been consistent.See, Meltzer and Richard (1981).
There is an open discussion about the relationship between globalisation and inequality.Among the authors who argüe that globalisation reduces inequalities are Williamson (199S), Lindert and Williamson (2001), Dowrick and DeLong (2001) and Wei and \Vu (2001).For a review of those authors, see making process, the influence of national pressure groups will also be reduced and, consequently, the demand for government gro\vth will also diminish.
The balance, however, seems to in favour the main prediction mentioned above.Indeed, classical research in the fíeld such as Peltzmans' (1980) reaffírms the idea that government growth is related to income differences.Although Pelzmans' empirical fíndings, which sought to establish a relevant influence of the growing cohesión of the middle classes on government growth in the US after the Second World War, are not conclusive at all, his hypothesis would allowus to conclude that if globalisation increased job insecurity, it would contribute to a decline of the political influence of the middle classes and the consequent reduction of government growth.Olson (1982), in tura, argües that there is a cióse correlation between longperíods of political stability and the emergence of political pressure groups, which would tend to positively influence public spending.Finally, the Leviathan theory assumes that since the state monopolises power, public spending might increase with greater centralisation and decrease with greater fragmentation.Then, if globalisation, as many argüe, produces a loss of state monopoly power and the fragmentation of its authority, it would give rise to a consequential decrease on public spending [Brennan and Buchanan (1980);Nelson (1987); Zax (1989) and Gates (1989)].
It is a basic premise that globalisation can genérate a more transparent taxsystem.
Finally, it has been argued that government growth is facilitated by fiscal il-lusion37.However, if we accept that globalisation may affect not only the tax base but also tax structures, fiscal ¡Ilusión may have been similarly affected.Indeed, it is a basic premise that globalisation can genérate a more transparent national tax system but, at the same time, a less progressive income-tax schedule.If this were to be the case, fiscal illusion would be reduced, and government growth subjected to a downward pressure.

Empirical evidence for globalisation and public spending
Let us consider some of the empirical evidence of the effects of globalisation on public spending.Garrett (1995) and (1998), Cusack(1997)íSwank (1997), Quinn(1997),andRodrik(1997); (1998) study the impact of a given measure of globalisation on public spending, mostly in OECD countries.However, Richardson (2002), pp. 22 to 26. In tura, Quinn (1997), Rodrik (1998), andWade (2004) fínd a positive relation between globalisation and inequality.Fiscal illusion refers ío a sort of a special status that the tax price of public ¡nput would have, by which voters would tend to underestimate it.This special status, in turn, would allow governments to increase taxes of public input without any significant political cost.See Oates (19SS).some of these authors follow specifíc lines of research.
A study by Cusack shows a significan!negativa correlation between capital marketintegration and government growth.Garrett (1995), Cusack (1997) and Swank (1997) adopt a political economic model.Specifically, Garret (1995) and Cusack (1997) analyse to what extent the margin for government discretion has been reduced by globalisation and the role of political party orientation in the determination of redistributionpolicies.Swank (1997), in turn, focuses on the role of democratic institutions as mediating actors between globalisation and national policy responses.By contrast, Quinn (1997), Rodrik (1997) and(1998), and Garrett (1998) assume an apolítica!perspective.
a) Empirical research based on the political-economic model.Garrett (1995) uses a percentage of GDP as a dependent variable government spending and, economic growth, unemployment, trade shares and an Índex of capital mobility as independent variables38.He fmds a negative correlation between capital-market and trade integration on one hand, and the level of public spending on the other.Additionally, he establishes that there is no signifícant positive correlation between the índex «Left labour power» and public spending, considering those variables alone.However, the interaction between a high level of capital market and trade integration with high scores of the índex of «Left labour power» appears to be positively correlated with public spending.On the contrary, at a low level of capital market and trade integration with similar scores of the índex of «Left labour power», government spending varíes negatively [Schulze andUrsprung (1999): 334], Cusack (1997) measures the impact of capital market integration on public spending on the basis of data for OECD countries for the years 1955-198939.The main results of his study show a signifícant negative correlation between capital-market integration and government growth.In turn, the political stance of the cabinet is negatively correlated with government spending when the political tendency of the cabinet is to the Right, and positively associated in the case of the See note 19.He also includes as independent variable an Índex that captures the centre of gravity of government and legislatura on a íeft-right scale.Additionally, he considers an Índex of left-labour power andan Índex measuring the concentration of trade unionism.See Garrett( 1995), p. 659.Cúsale uses the Feldsíein/Horioka measure, \vhich assesses the absolute valué of the normalised difference between prívate saving and investment.He also uses the notion of change in non-defence government outlays as a percentage of GDP.Additionally, as well as Garrett (1995), Cusack includes an Índex to capture the centre of gravity of government and legislature on a left-right scale.He uses two explanatory variables: the centre of gravity of the legislature and the distance beíween the centres of gravity of government and legislature, see Schulze and Ursprung (1999), p. 330-333.Left.He also discovered that the difference between the political stance of the cabinet and legislature is positively correlated with government spending.Swank (1997) measures the impact of financial market integration on government spending and the interaction of those variables with the role of democratic institu-tions40.He fmds that while capital market integration (with the exception of Foreign Direct Investment) does not signifícantly influence government growth, trade integration does have a positive influence on government growth.He also discovers that capital market integration shows a signifícant positive influence on that growth in countries characterised by high corpora-tism41, high consensus democracy42, and low dispersión of authority43.On the contrary, it has a significant negative influence on government growth in countries characterised by low corporatism, low consensus democracy and a high dispersión of authority44.

b) Empírica! research based on the apolítica! inodel
Quinn (1997) examines the impact of capital mobility on government expenditure on the basis of data from 38 coun-tries45.His main fínding is that capital mobility is positively correlated with government spending.Rodrik (1998), in turn, studies the impact of openness46 upon government spending47 frorn data collected from over one hundred countries48.He also includes terms of trade volatility as an explanatory variable.His main conclusions are that trade integration is positively associated with government spending and that the interaction between terms of trade volatility and openness show a highly signifí-To capture the degree of capital market integration he uses a measure of total inflow and outflow of capital, foreign direct investment and total borrowing on International capital markets as a percentage of GDP.He also utilises an Índex based on the IMF classification of restriction on cross-border movemenís and the absoluto valué of covered interest pariíies to capture the degree of capital market liberalisation.Additíonally, he includes the standard measure of trade integration, that is, exports plus imports divided by GDP.It also includes the total public sector outlays as a percentage of GDP.For example, Norway and Sweden.For example, Belgíum and Netherlands.For example, Denmark and Finland.Swank (1997) reaffírrns his results in a second more specifíc regression, which employs public consumption instead of total public sector outlays.For the rneasure of capital mobility and the individualization of the 38 countries see Note 21, p. 12. Quinn (1997) uses a measure of government outlays net of defence and education expenditures as a percentage of GDP and an addítional measure of government welfare and social security payments, see Quinn (1997), p. 537.He uses the exports plus imports divided by GDP to measure trade integration: openness.Data is taken from world tables 5.6.See Rodrik (1998), p. 999.He uses government spending as a share of GDP.The data is taken from the World Bank's World Data 1995.See Rodrik (1998), p. 999. Rodrik (1998) considers data from developed and developing countries, including OECD, Latín American, East Asia, (ex) Socialist and Sub-Saharan África countries.See Rodrik (1998Rodrik ( ), p. 1003. .cant positive impact on government consumption as a share of GDP.
Most poorer countries face high trade risks and therefore their public spending varíes positively with openness.
In his 1997a book Rodrik uses a narrower datábase than ¡n his 1998 study.In fací, in the former he only includes OECD countries in order to analyse the impact of openness and terms of trade volatility on social security and welfare expenditure.His main fmdings are that openness and terms of trade volatility have a negative effect on social security and welfare expenditures; the interaction of openness and terms of trade volatility indícate that countries with high levéis of openness and trade volatility present larger amount of welfare spending than countries with low levéis of openness and trade volatility, and that openness has a particularly negative strong effect on welfare spending when there is a high level of capital mobility.These somewhat mixed results, compared to the Rodrik's (1998) findings, have been expíained by the size and composition of the data utilised in both studies.Indeed, the positive correlation between openness and public spending foundbyRodrik (1998) could simply be due to the fact that this research considers a broader sample of countries, including relatively poor countries.Most of these face high terms of trade risk and therefore their public spending varíes positively with openness.On the contrary, OECD countries, which are included in Rodrik's earlier study (1997a), as a general rule present lower levéis of terms of trade risk and thus their public spending is negatively correlated with openness.
Finally, Garrett (1998) seeks to measure the influence of trade integration49 and capital mobility50 on government con-sumption51.Additionally, he includes an Einteraction terms' variable52 to capture the influence of globalisation on the convergence of policy regimes among nations.The main results of Garrett' s study (1998) are that trade integration is positively associated with government spending but there is not any significant influence of fínancial and real capital mobility on government consumption.Moreover, with the 'interaction terms' variable trade integration generates a tendency of divergence in policy regimes of the countries analysed.
The Globaüsation Debate Revisited

An assessment ofthe empirical evidencefor globalisation andpublic spending a) Empirical research based on the political-economicmodel.
As a general rule, the fmdings of the three works analysed conclude that the size of govermnent is not necessaríly reduced in a globalised era and that ideological factors as well as social structure differentiations play a key role in determining íhe level of government expenditure.
Indeed, Garrett's fmdings (1998) support the hypothesis that governments that are ruled by strong Left parties tend to increase their public spending with a high level of globalisation.Cusack (1997), in turn, concludes that the rise of globalisation has not eliminated the influence of politics.In fací, the Left seeras inclined to increase the size ofthe public sector and the Right to reduce it [Cusack (1997): 392]53.Finally, Swank (1997) concludes that political institutions and social structures play an important role in mediating the impact that globalisation might exert on the government growth.

b) Empirical research based on the apolitical model
The three cross-section studies that we have examined show that there is an important correlation between globalisation and government spending.Indeed, all three fmd that the demand for social insurance programmes grows with the increase of international economic integration.Howẽ ver, as they all use different methodologies to capture the degree of international economic integration, their results, which in many senses are contradictor/, must be treated with caution54.
Empirical evidence points to a positiva correlation between globalisation and government growth.
To summarise, the empirical evidence analysed on both models seems to lend substantial support to the 'compensation hypothesis', since not only would globalisation be positively correlated with government growth, but also suggests, as predicted, that governments are not very vigorously cutting backthe provisión of goods and services to those on whom they rely for political support.
However, one of the main methodological objections of this conclusión is that Cusack (1997) points out that although the integration of states into the international capital market has, in fact, had a depressing effect on the size of public spending, there is no evidence to conclude that these trends have ovenvhelmed the relevance of domestic partisan political influences in the determination of public expenditure.See Cusack (1997), p. 392.This is particularly notorious between the works of Quinn (1997) and Garrett (1998).In fact, while the former fínds a positive association between capital mobility and government, the latter shows the ¡nverse tendency.
none of these works have adequately taken into account that government spending is a 'mixed bag' rather than a single variable, composed of diverse Ítems, most of which react in different ways to the constraints imposed by globalisation.Therefore, it is relevant to study separately the impact of globalisation on the structure of public spending.

International competition for productive factors generates positive
externalities for immobile factors.

Globalisation and the structure of public spending
It has been argued that the analysis of the influence of globalisation on the different components of public spending is the key factor to assess the 'compensation hypothesis' [Schulze and Ursprung (1999): 337].Particularly, if we assume that the increase of some types of government expendiíure, such as public infrastructure spending, it would support the efñciency hypothesis rather than the compensation hypothesis [Aschauer(1988)].
Additionally, the analysis of the structure of public spending is methodologically relevantbecausethe influence of openness on government growth could be a consequence of country size effects instead of trade integration.Since government con-sumption as a percentage of GDP is negatively correlated with population size (given economies of scale and a fixed cost of government service) and openness also varíes negatively with population size, government consumption could vary positively with openness, even if there is not a direct causal link between both variables [Schulze and Ursprung (1999), p. 337].
Therefore, the real variables to be analysed should be social insurance components of government spending in particular, instead of government spending in general.

Theoretical links between globalisation and the structure of public spending
International competition forproductive factors, particularly capital competition, brought about by globalisation, tends to genérate positive externalities for immobile factors.Inthis sense, in aglobalised scenario we expect to fínd increased government spending in favour of these factors, especially in terms of public goods provisión which improves theirproductiv-ity55.In turn, the increasing economic instability and uncertainty often associated with globalisation tends to produce an increased demand for redistribution of programmes.Therefore, globaíisation would also be positively associated with redistributive spending programmes [Schulze and Ursprung (1999): 338].
Examples of these public goods provisions are: public infrastructure, training labour programmes and improving environment and cultural conditions.See Keen and Marchand (1997).

Empirical research on globalisation and the structure of public spending56
Most empirical studies that have tested the theoretical links existing between globalisation and the structure of public spending have used either the economic classifícation or the functional classification of government expenditures57, and have been classified within either political economic models orapolitical models.
a) Empirical research based on the política l-economic model Hicks and Swank (1992) have studied the influence of political orientation of the competing parties on welfare state expenditure58 and have included open-ness59 as an explanatory variable of welfare state expenditures.They fínd that there is a positive and signifícant córrelation between openness and social security benefíts spending as well as a direct proportional influence of Centre orRight opposition parties upon the decrease of welfare expenditures of a Left government.They also discover a direct proportional influence of Left opposition parties on the increase of welfare expenditures of Right and Centre governments.
Huber, Ragin and Stephens (1993) also assess the influence of partisan political orientation and openness60 on welfare state expenditures61.They find that there is no positive and signifícant correlation between openness and social security benefíts spending, but that such correlation exists between openness and social security transfers.Additionally, they find that both Social and Christian democratic parties have a positive influence on welfare state spending, with Social democratic parties having a stronger impact than Christian democratic parties.
Openness and social security benefíts are positively correlated.
However, it is worth noting that empirical quantifí catión of the theoretical premises of the relationship between globalisation and the structure of public spending is particularly complicated to assert because it is difficult to distinguish public expenditures motivated by effíciency considerations from those motivaíed by involuntary redistribution.See Schulze and Ursprung (1999), p. 338.The economic classifícatión distinguíshes five different categorías of government expendiuire: (í) government consumption (purchases goods and services, (ii) current transfer (redistribution programmes), (iii) capital formation (public investment), (iv) interest payments on government debí and (v) capital íransfer.The functional classifícation, in turn, comprises of four policy áreas: (i) the traditional dornain (general public services, public order, safety and defence) (ii) the welfare state (education, heath, social security welfare and housing) (iii) the mixed economy (economic services) and (iv) other functions.They capture the welfare state expenditure by using an indicator of social security benefíts taken from ILO (International Labour Organization).Exports plus imports divided by GD.R Exports plus imports divided by GDP.In this case, the dependent variable of welfare state expenditures is the measure of social security transfer payments used by the OECD.Swank (1997) measures the influence of capital market iníegration62 on the share of outlays for social transfers as a percentage of GDP.He finds that capital market integration does not present a signifícant negative influence on social transfers.However, fínancial integration63 presents a positive influence on the volume of social transfers and capital market integration has a signifícant positive influence on social transfers in countries characterised by high corporatism, high consensus democracy and a low dispersión of authority.
b) Empírica!research based on the apolitical model.Quinn (1997) analyses the impact of capital mobility on government welfare and social security payments.He concludes that capital mobility generates an increase in welfare and social security payments.Garrett and Mitchell (1997) investígate the impact of globalisation64 on various types of income transfer programmes in OECD countries.They fmd that total trade does not have a signifícant positive influence upon any category of public transfers.In turn, trade exposure multiplica with terms of trade volatility only, produces a signifícant positive influence on unemployment benefits expenditures.In the case of the volume of imports coming from low wage countries, they fmd a signifícant positive influence on total income transfers, old-age pensions and unemployment benefit expenditures.Finally, Garrett and Mitchell (1997) show that fínancial market integration presents a signifícant positive effect upon total income transfers, family allowance and 'other transfers'65.
Partisan characteristics would still be important in determining the patterns of public spending.Rodrik (1998) has studied the impact of openness on different types of government spending and finds that openness has a signifícant positive influence on most of them.However, this positive correlation is truly signifícant in the 1990-1992 regression, but not in the 1985-1989 one.Additionally, the only Ítem of government spending in which openness does not exert a signifícant positive influence is on interest payments on public debí.
In an earlier study Rodrik (1997a) analyses the influence of openness and terms of trade volatility, social security and welfare expenditures for OECD countries.In that work he fínds that openness He captures the degree of capital market integration by the following measures: total capital flows, FDI, borrowing and capital liberalizaron.Thís variable is captured by measuring the inverse of covered interest rate differentials.Globalisation is captured by three measures: (i) The volume and volatility of trade (ii) the volume of imports coming from low wage countries and (iii) financia!market integration, which is measured by covered interest rate differentials.«Other transfers» include benefits for sickness and disabilities.and terms of trade volatility separately exert a negative effect on social security and welfare expenditures.He also finds that the interaction of openness and terms of trade volatility indícate that countries with a higher degree of openness and exposure to substantial external risk produce higher welfare expenditures than tliose countries that are open but less exposed.
Additionally, the results ofempirical research based on a political-economic model show that partisan characteristics are still important in determining the pattems ofpublic spending [Hicks and Swank (1992)].In fací, the design of the democratic institutions that regúlate the interaction between governments and economic interest is central to defming the specifíc way in which globalisation affects the structure ofpublic spending [Huber et al, (1993)].Furthermore, in countries with social corporatism, consensus democracy and centralisation, the compensation effect of globalisation is relatively stronger than that of the efficiency effect of globalisation [Swank, (1997)].
In ten, the fmdings ofempirical research based on apolitical models establish a positive relationship between globalisation and welfare state spending.However, this correlation is more explicit and constant between capital mobility [Quinn, (1997)] or financial market integration [Garrett and Mitchell, (1997)] and different Ítems of welfare state spending than in the correlation of openness and welfare state spending, in which results are not totally consistent.Indeed, while Garrett and Mitchell (1997) do not fmd any signifícant positive association between openness and welfare state spending, in his two studies Rodrik apparently fmds contradictory results.In fací, in the more recent one he reports a signifícant positive influence that contrasts with the earlier one, where he finds an inverse result.However, the apparent contradiction is explained when the variable of'openness' is combined with terms of trade volatility.In this case, Rodrik (1997a) shows that the interaction of openness and terms of trade is positively correlated with welfare state expenditures.Finally, Rodrik's 1998 and1997a studies agree with the statement that in countries with a simultaneous high level of openness and high exposure to substantial external risk, welfare state spending varíes positively with openness.This result is coincident with the final conclusión reached by Garrett and Mitchell (1997) who fmd a positive significant influence of both openness and terms of trade acting together on unemployment benefit expenditures.
In a global economy states have neither beconieunnaturalbusiness units ñor have the expanded their policy choices.

CONCLUDING REMARKS
Paraphrasing the question raised by Dani Rodrik: has globalisation gone too far?we must conclude that the answer, at least in the field of fiscal policies in OECD countries, to which the examined studies are mostly restricted, is far from uniform and conclusive.One thing that however seems to be clear ¡s that states have neither become unnatural, even impossible business units in a global economy, as suggested by Ohmae (1995): 5, ñor have they necessarily experienced an expansión of their policy choices, as implied by Weiss (1998): 184.States are clearly not disappearing from the International arena but they do seem to be affected by 'new' constramts that diminish their capacity to autonomously setup both tax and government expendíture policies.
Notwithstanding, it is worth noting that the 'autonomous character' of fiscal policies is here understood as the capacity that governments once had in formulating their policies, to only accept pressures coming from traditional domestic factors; a reduc-tion of such autonomy being a situation in which governments seemto be compelled to accept a greater influence of global factors such as openness, capital mobility or high trade volatility.If this is the case, in the context of globalisation a reduction of the autonomy of fiscal policies of states has clearly takenplace.
Atotally different conclusión emerges, however, if we assume that these constraints, that are affecting the autonomy of governments to setup their fiscal policies are exclusively due to globalisation understood as an increasing process of openness and capital mobility, which acts as a natural causal phenomenon.In other words, if there is really a myth about globalisation, it is that which affirms that £Politics5 is no longer relevant for the configuration of the new form of the global economy.The (ideological) iliegitimate underlying assumption in the debate of the impact of globalisation upon fiscal policies is that globalisation is a type of natural phenomenon, that leaves players with no other option but to accept it as inevitable.Indeed, apart from the role that states might have played in the origin of globalisation, as has been highlighted by Gilpin (1987), some of the most outstanding studies on fiscal policies in a globalised era are those which show that constraints imposed by globalisation are not only far from being conclusive but also that the real impact of these constraints depends on the type of society, government and level of development that each country presents [Garrett (1998); Cusack (1997); Swank (1997); Hicks and Swank (1992);Huber etal, (1993)].
The most accurate conclusión about the impact of globalisation on fiscal policies of the states, though also the most challenging one, would thus be that there is no conclusión at all if we understand globalisation as a final paradigm that governments unavoidably have to adopt.Therefore, far from being an inevitably constraining phenomenon regulated according to 'economic invariable laws', globalisation turns out to be, betterunderstood if it is observed as an economic political process (historically determined), as classic political economy has ahvays insistently emphasized.
Quinn (1997)utward current account transaction are coded Curren t.The sum of íhe six dimensions of Current and Capital generales a 0-12 score, ranging from most closed (0) to mosí open (12) economy.Additionally, there is a seventh dimensión which captures international legal agreements (Agree) that constrain a nation's ability to restrict exchange and capital flows.SeeQuinn (1997), p. 535.21 TJie dependent variables are the corporate tax revenue that isconsidered; (i) as a percentage of GDP (ii) as a percentage of individual tax revenue and (iii) as a percentage of total government tax revenue.The last two measures analyse redistribute aspects of tax collections.Data covers 38 countries between the years 1974 and 1989: Latín American Nations: Bolivia, Brazil, Chile, Costa Rica, 22 Domestic variables includes percentage changes on (i) ínvestment, (ii) real net operating income, (iii)