Sage Journals: Discover world-class research

Abstract

How does regionalization affect national social policies? Although there is an extensive literature on the effects of globalization on social protection, the literature on the impact of regional integration is much less developed. I argue that the distinctive nature of regionalization processes calls for rigorous empirical testing of the domestic policy effects of regional integration. To this end, using an innovative dataset that measures the degree to which countries are integrated into regional economic and political organizations, this article uses statistical analysis to consider the influence of regional integration on government social spending. The results are surprising: regionalization has a significant and positive relationship with government social spending, controlling for other factors, even when the European Union countries are excluded from the analysis. In fact, in the EU countries increasing regionalization is associated with lower social spending levels. These results suggest that regional economic and political integration does not necessarily lead to a “race to the bottom” of social spending. Instead, regionalization appears to accommodate wide divergence in national social policy commitments.

Keywords

Globalization race to the bottom regional integration regionalization social policy

Introduction

How does regional integration affect national social policies? Despite extensive literature on globalization’s effects on social spending, and despite dramatic growth in the number of regional organizations and the intra-regional movement of goods, capital, and labor, the literature investigating regional integration’s impact on social policies is comparatively underdeveloped. Furthermore, given that most international trade is regionalized, rather than globalized, some scholars argue that regional integration’s effect on national policies may be stronger than the effect of globalization (Beckfield, 2006).¹

This article presents an empirical analysis indicating that increasing regional integration is associated with increased levels of social spending. Its findings show that as states integrate into regional political and economic institutions, they do not necessarily face a “race to the bottom” in social spending in which governments compete to cut spending—and the tax burden on domestic firms—in order to attract foreign firms and prevent domestic firms from relocating to neighboring states where it is cheaper to do business. Instead, I find that in most states, regional integration is associated with higher levels of social spending. The research makes several contributions. First, it contributes to the nascent literature on social policies in the developing world.² Second, it applies well-developed and oft-tested theories about the domestic effects of international economic pressures to cases outside of the OECD context. Scholars writing in the “race to the bottom”/ “climb to the top” school either have assumed these theories would not hold or have neglected to test these theories beyond OECD states. Finally, I employ an innovative measure of regional integration developed at the regional level of analysis by other researchers (Efird and Genna, 2002), applying it to individual countries from 1980–2000, a period of time that captures important changes in levels of regional integration all over the globe.

Regionalization, globalization, and social spending

Regional integration refers to the process of increasing political and economic cooperation among states in close geographic proximity to each other.³ Scholars often conceptualize regional integration as two broadly conceived processes: economic integration and political integration. Economic integration, they argue, is a more informal, societally-driven process that removes trade and investment barriers (Haggard, 1993; Katzenstein, 2005). This type of integration has been occurring at significant levels around the world. Political integration is a more top-down, state-driven process of institutional creation at the regional level (Fawcett, 2004; Katzenstein, 2005). Marchand et al. (1999) conceive of political integration as a policy project motivated by interests, but also by identity and ideological factors. Political integration, including the creation of regional political institutions, is a process that has developed much more slowly outside of Europe, although recent developments in Southeast Asia and South America suggest renewed efforts on this score (Börzel and Risse, 2009; Jetschke, 2010). Over the last few decades, the degree of integration that regional trading organizations have actually implemented has increased substantially (Efird and Genna, 2002; Haftel 2007). Figure 1 demonstrates these trends.

Figure 1.

Levels of regional integration.

Regional integration is contested because, like globalization, it creates national policy externalities as labor and capital mobility increase. Certainly, regional integration is related to the larger process of globalization. It is, however, a distinct form of international embeddedness that differs from globalization in several important ways. Regional integration can place different demands on governments than globalization processes more generally because it can be more strongly institutionalized, can generate higher cross-border flows, and can take on elements of political integration or supranationalism that are absent from the anarchic global level.

Economically, states may pursue regional integration in order to achieve fuller liberalization than they would be able to achieve under multilateral trading agreements (Lawrence, 1996). With greater capital mobility within the region, states may have an incentive to cut social programs and corporate tax rates in order to make their economies more attractive to businesses. Social dumping and migration to member states with more generous social policies can also put strain on those welfare states. Beyond economic pressures, regional integration includes a political dimension that exists to a much lesser degree at the global level. Regionalization is more highly institutionalized than globalization and it can lead member states to pursue certain policies in a way that most global organizations, including the World Trade Organization, are unable to do (Beckfield, 2006). For example, Pierson (1996) argues that, in the EU member states, national political actors can pursue welfare state retrenchment more successfully than they could prior to EU membership because they can blame their cuts on the EU and suffer less political backlash from voters.

On the other hand, regional integration could have positive social outcomes if states compensate society for economic insecurity with increased social spending, or if social policies are created or protected at the supranational level. Caporaso and Tarrow (2009) argue that this is exactly what has happened in the EU. Examining the implications of the free movement of labor in the EU on national social policy frameworks, they contend that the European Court of Justice has actively promoted the protection of social rights at the regional level (though they did not look at spending). Finally, regional integration may have no effect on social policies, allowing for continued divergence in these policies cross-nationally.

How has social spending changed over the past few decades, and how can we link these changes to globalization or regionalization processes? Missing data is a significant problem for assessing spending trends in non-OECD countries. The data on spending across the developing world is extremely sparse, so we should be very cautious when drawing conclusions from the data that is available. The graph in Figure 2 presents spending averages in OECD and non-OECD countries over the last few decades; the trend for non-OECD countries especially should be interpreted carefully, given the problems with missing data.⁴

Figure 2.

Social spending as a percentage of total government expenditures.

Theoretical perspectives

Existing theories of both globalization and regionalization can shed light on the ways regional integration may affect national social policy spending. “Race to the bottom” or convergence approaches would expect regionalization to exacerbate the economic pressures placed on states by greater economic interdependence and globalization. Scholars argue that the pressures to be competitive in global markets reduce national policy autonomy in areas ranging from social policy to fiscal policy to monetary policy (Cerny, 1995; Rodrik, 1998). Some analysts argue that the “race to the bottom” effect is most likely to operate on developing countries (Rudra, 2002). If increasing market integration pressures governments to adopt similar policies, we should expect the integration of regional markets, which is often deeper than integration achieved by global institutions like the WTO, to heighten these pressures. Regional integration could lead to cuts in social spending as countries within regional organizations compete to attract (or keep) capital. By reducing expenditures, governments may be able to lower corporate tax rates, making their country a more attractive place to do business. The “race to the bottom” hypothesis suggests that as regional integration increases, especially in terms of trade and capital openness, social spending is likely to decrease.

Other scholars argue that the insecurities imposed on social groups by global competition fuel demands for states to increase social spending. These demands produced the post-WWII expansion of the welfare state in most of the advanced capitalist democracies (Cameron, 1978; Katzenstein, 1985). Under conditions of regionalization, this welfare state expansion can be viewed as a way for governments to compensate workers for the stronger wage competition of the regional market’s larger labor pool. The compensation hypothesis suggests that as regional integration increases, social spending is also likely to increase.

Finally, the null hypothesis is that regional integration is not related to changes in national social spending. If we do observe cuts or growth in social policy among members of regional trading agreements, these changes will be the result of other processes, such as globalization or domestic changes. Many argue that changes in welfare state spending are primarily the result of domestic processes such as deindustrialization or demographic changes rather than international-level processes such as globalization or regionalization (Huber and Stevens, 2001; Iversen and Cusack, 2000; Mosley 2003). I test some of these domestic-level variables in this study.

Variables, data, and methodology

In this study, I use time-series cross-sectional panel data for all possible countries, from 1980–2000, the years for which the most data is available. Despite the significant problems with missing data for non-OECD countries on several variables, including my dependent variable, a major purpose of this study is to understand the effects of regionalization in a wider comparative context that goes beyond the EU and other OECD countries. Because of sparse data from the developing world, however, results should be interpreted cautiously.

The unit of analysis in this study is the country-year. The dependent variable is social welfare spending. I measure this as the ratio of social welfare spending to total government expenditures.⁵ This measure may better capture government spending priorities than does a measure of social spending as a percentage of GDP (Wibbels and Ahlquist, 2011a: 138). Social welfare expenditures, as defined by the IMF measure, include government spending for welfare purposes, including pension and retirement benefits, unemployment benefits, sickness and old age benefits, and family allowances. I draw data for this variable from Wibbels and Ahlquist (2011b), whose data is sourced from IMF Government Finance Statistics, the Economic Commission of Latin America (CEPAL), and from their own primary data collection efforts.

The key independent variable is level of regional integration. To measure the extent to which a country has integrated into a regional body, I use Efird and Genna’s (2002) Integration Achievement Score (IAS). Efird and Genna developed this multi-dimensional measure from original data on the levels of regional integration actually achieved in all cases of regional integration reported to the WTO, up until 2004. They assign each organization an IAS score for each year that the organization has been in existence. Their coding system looks at six categories of regional economic and political integration—trade, capital mobility, labor mobility, supranational institutionalization, monetary policy coordination, and fiscal policy coordination—and gives each regional organization a score between zero and five for each category. Each organization’s IAS score is the average of the levels of integration achieved in each category. The coding system is described in full detail in the appendices at the end of this article. The IAS measure is one of only a small handful of attempts to code regions according to the level of integration that is actually implemented.⁶ In this way, this measure should allow us to advance studies of regionalization and empirically test theories of regionalization’s causes and effects.

In this study I have applied this measure of regional integration to the country level of analysis, although Efird and Genna devised the measure with the region as the unit of analysis. For example, they measured the level of regionalization achieved by Mercosur for each year of Mercosur’s existence. Because I am interested in the domestic effects of individual countries’ levels of integration into regional frameworks, my unit of analysis is the country-year. I have thus coded each member state of a regional organization with the IAS score that the organization achieved in that year. A potential problem with this application of the IAS score is that there may be significant variation in the degree to which member states of the same organization implement agreed-upon integration measures. There may also be significant cases of noncompliance. To address this, I have adjusted each country’s IAS score in a given year to take into account any formal ways it opted out of deeper integration that other member states of the organization pursued. For example, the EU countries that have opted out of the single currency receive a lower score than those that have adopted the euro. While this partially addresses the issue of variation in regionalization among members of the same organization, this measure cannot account for informal ways that individual member states may obstruct regional integration.

Finally, I limited the cases of regional integration in this study to multilateral organizations that have achieved at least a free trade area: the EU, Mercosur, the Association of Southeast Asian Nations (ASEAN), the North American Free Trade Agreement (NAFTA), the European Free Trade Area (EFTA), the Central American Common Market (CACM), the Andean Common Market (ANCOM), and the Caribbean Community (CARICOM). I did not include members of regional organizations in Africa or the Middle East in this analysis because of severe missing data problems for my dependent variable. I assigned countries a score of zero in all years prior to joining a regional organization. When a country was a member of more than one regional body, I assigned it the higher of the two scores.

I have included several variables in my model to control for other commonly theorized determinants of social welfare spending. First, it is important to attempt to distinguish between regionalization and globalization effects, though these certainly overlap. To control for the influence of globalization, I control for trade openness, measured as the total number of imports plus exports as a percentage of GDP (“World Development Indicators,” World Bank), and for FDI (foreign direct investment) openness, measured as the total stocks of inward plus outward FDI as a percentage of GDP (“UNCTADStat,” United Nations Conference on Trade and Development). Second, I control for several domestic-level variables that have been theorized to affect social spending levels. I control for GDP per capita, the percentage of the population that is over 65 years old, the percentage of the population working in the service sector (all drawn from the World Development Indicators), and the strength of democracy (from the Polity IV score, Marshall et al., 2004). These domestic variables test arguments that higher levels of development, democracy, deindustrialization, and population ageing should put upward pressure on governments’ levels of social spending.⁷ Finally, I control for EU membership in order to ensure that the “European model” of generous national welfare states under high levels of regional integration does not bias the results of my broader cross-national model.

I estimate the effects of regional integration on social spending for the time period of 1980–2000. With cross-national data that spans such a broad swath of countries, there are many country-specific factors that could influence social spending levels. In order to account for these country-specific factors, and for omitted variable bias, I estimate the model as a time-series with country fixed-effects. While my baseline model includes all countries, I run an additional model that excludes EU countries, as well as a third model consisting of only EU countries. Given the emphasis in the literature on the “uniqueness” of the EU model of high social spending combined with high regionalization, it is important to ensure these countries do not bias the results of the broader cross-national models.

Results

In Table 1, for the full sample of countries (Model 1), regional integration has a significant and positive effect on social spending, suggesting that in a given country, as integration increases over time, social spending as a percentage of total government expenditures increases as well. Substantively, a one-point increase in a country’s IAS score is associated with a 1.9% increase in social spending as a percentage of total government expenditures. The only other independent variables that reach statistical significance in this model are GDP per capita and the dummy variable controlling for EU membership. Higher levels of economic development are likely associated with higher social spending because as economies develop, domestic social groups become better mobilized and acquire more resources to lobby governments successfully for welfare benefits.

Table 1.

Social spending as a percentage of total government expenditures.

	Model 1	Model 2	Model 3
	(All countries)	(Non-EU countries)	(EU countries only)
	Time-series	Time-series	Time-series
	fixed-effects	fixed-effects	fixed-effects
IAS score	1.872*** (0.542)	1.638*** (0.557)	−2.048** (0.822)
Trade openness	−0.014 (0.015)	−0.023 (0.015)	0.006 (0.056)
FDI openness	0.011 (0.015)	−0.059*** (0.021)	0.023 (0.022)
GDP per capita (ln)	3.303*** (0.691)	1.415 (0.812)	4.169*** (1.429)
Polity	0.002 (0.072)	−0.033 (0.067)
Population > 65 years	−0.830 (0.498)	3.069*** (0.727)	−2.960*** (0.801)
Service sector employment	−0.014 (.036)	−0.044 (0.033)	0.277 (0.191)
EU dummy	−10.243*** (1.413)
Constant	5.037 (5.005)	−9.445 (5.342)	27.353*** (11.228)
Number of observations	528	346	182

p < 0.1, **p<0.05, ***p<0.01.

The results of the full model suggest that, as countries become more integrated into regional economic and political structures, they will not necessarily face economic pressures that force cuts in social welfare spending. In fact, countries that are more economically open and politically integrated within their region tend to spend more on social welfare. But considering that that level of development and EU membership are the only other two independent variables that predict social spending at a statistically significant level, we might reasonably wonder if the highly integrated EU states are driving these results.

To address this concern, Model 2 (Table 1) employs a time-series fixed-effects analysis on a sample that excludes EU countries. In case the effects of EU candidacy may also bias the results, I also excluded states that would become EU members in the future. The results of Model 2 again indicate that as a country becomes more regionally integrated over time, its social welfare expenditures are likely to increase as well, even after controlling for the effects of increased global integration (captured by the trade and FDI openness variables) and for domestic demographic and structural factors. The exclusion of the EU countries increases confidence in the general applicability of this relationship.

Two other variables are statistically significant in this model: FDI openness and size of the elderly population. FDI openness has a negative effect on social spending, though its effects are substantively very small. Countries with substantial levels of foreign investment may be pressured to keep corporate taxes low, and this may constrain government social expenditures. Not surprisingly, having a large population that is over 65 years old is positively and significantly associated with social spending. The relatively large size of this coefficient gives some support to arguments that domestic demographic have a greater influence on social spending levels than international factors (Iversen and Cusack, 2000; Pierson, 2001).

Finally, I estimated the effects of increased regional integration on the EU member states only. Again using time-series fixed-effects, the results were surprising. I found that increasing levels of integration over time are associated with lower welfare spending as a percentage of total expenditures in these states. This finding is the opposite of what I found in the full sample as well as in the sample consisting of only non-EU countries. This finding suggests that in EU countries where social expenditures already make up a significant percentage of government expenditures, openness to the well-developed and highly competitive European markets does constrain social spending. The fact that neither openness to trade nor FDI are significant suggests that there is something specific about integrating into EU structures, rather than global structures, that constrains social spending. This could be related to monetary integration in the EU, to the constraints on national fiscal policy imposed by the Maastricht criteria, and to the limits on member states’ public debts and deficits. The democracy variable is excluded from this model, as slow-moving institutional variables tend to be highly collinear with country fixed-effects, and there is extremely little variation on the Polity score for the EU sample in this time period. There is also little variation in the percentage of the population over age 65 in this model, so the large, negative, and significant coefficient on that variable may not be reliable.

Interestingly, trade openness is not significantly related to social spending in any of the models, nor does its coefficient have a substantial size. Social expenditures are better predicted by the degree to which countries are more broadly integrated into economic and political institutions in their region. Trade is certainly a component of this regional integration, but other dimensions of regionalization, such as greater labor mobility, may counterbalance potential “race to the bottom” competitive pressures.

As a robustness check, I estimated the three above models using an alternative indicator of regional integration developed by Haftel (2007) that combines measures of the scope of institutionalization and implementation of regional integration agreements. This measure was coded at five-year intervals, so there are considerably fewer data points. With the exception of the final EU-only model (Model 3, Table 1), for which there are too few observations for a meaningful analysis (n=24), the other two models (Models 1 and 2) generate similar results to those reported above. Level of regional integration is positive and significantly associated with social spending, even when controlling for other leading explanations in the international political economy literature.

Discussion and conclusions

This study uses one of a very few existing datasets on the implementation of regional integration to probe an understudied question, and it yields some interesting and unexpected results that provoke further analysis. Overall, the results of this study are consistent with theories of globalization or regional integration that predict continued divergence of national social spending levels. The results of my country fixed-effects models show that greater regionalization is associated with higher levels of social spending in the broad cross-national sample. However, a surprising finding was that this is not the case in the EU countries. Here, social spending decreased as regional integration increased, suggesting that regionalization does constrain social spending either when levels of integration are very high, or when social spending commitments are very high. The potential interaction between level of development and regionalization should be further explored.

Despite limitations introduced by missing data, especially for developing countries, this study is a first step in assessing how regionalization affects national social policy commitments outside of the EU. It is important to emphasize that the central finding that regionalization is associated with higher social policy spending in the broader comparative context, even when controlling for other international and domestic-level factors, should be carefully interpreted. The results here are supportive of the divergence hypothesis, rather than the convergence or “race to the bottom” hypotheses. I do not argue that regional integration causes governments to increase social spending. Yet, it is clear from this study that most countries that integrate into regional economic and political organizations will not be forced to reduce social spending as a percentage of government expenditures. Indeed, greater national commitments to social welfare often occur alongside regional integration.

Footnotes

Declaration of conflicting interest

The author declares that there is no conflict of interest.

Funding

This research received no specific grant from any funding agency in the public,commercial,or not-for-profit sectors.

Supplementary Material

The replication files are available at:

References

Adserà

Boix

(2002) Trade, democracy, and the size of the public sector: The political underpinnings of openness. International Organization 56(2): 229–262.

Avelino

Brown

Hunter

(2005) The effects of capital mobility, trade openness, and democracy in Latin America. American Journal of Political Science 49(3): 625–641.

Beckfield

(2006) European integration and income inequality. American Sociological Review 71(4): 964–985.

Börzel

Risse

(2009) Diffusing (inter-) regionalism: The EU as a model of regional integration. KFG Working Paper 7. Berlin: Freie Universität.

Brown

Hunter

(1999) Democracy and social spending in Latin America, 1980–1992. The American Political Science Review 93(4): 779–790.

Cameron

(1978) The expansion of the public economy: A comparative analysis. American Political Science Review 72(4): 1243–1261.

Caporaso

Tarrow

(2009) Polanyi in Brussels: Supranational institutions and the transnational embedding of markets. International Organization 63(4): 593–620.

Cerny

(1995) Globalization and the changing logic of collective action. International Organization 49(4): 595–625.

Efird

Genna

(2002) Structural conditions and the propensity for regional integration. European Union Politics 3 (3): 267–295.

10.

Fawcett

(2004) Exploring regional domains: A comparative history of regionalism. International Affairs 80(3): 429–446.

11.

Haftel

(2007) Designing for peace: Regional integration arrangements institutional variation, and militarized interstate disputes. International Organization 61(1): 217–237.

12.

Haggard

(1993) Comment. In: Frankel

Kahler

(eds) Regionalism and Rivalry: Japan and the United States in Pacific Asia. Chicago, IL: University of Chicago Press.

13.

Huber

Stevens

(2001) Development and Crisis of the Welfare State. Chicago, IL: University of Chicago Press.

14.

International Monetary Fund (2011) Government Finance Statistics. Available at: www.imf.org/external/data.htm.

15.

Iversen

Cusack

(2000) The causes of welfare state expansion: Deindustrialization or globalization?. World Politics 52(3): 313–349.

16.

Jetschke

(2010) Do regional organizations travel? KFG Working Paper 17. Berlin: Freie Universität.

17.

Katzenstein

(1985) Small States in World Markets. Ithaca, NY: Cornell University Press.

18.

Katzenstein

(2005) A World of Regions. Ithaca, NY: Cornell University Press.

19.

Lawrence

(1996) Regionalism, Multilateralism, and Deeper Integration. Washington, DC: Brookings Institution.

20.

Madeira

Caporaso

(2011) Regional integration. International Encyclopedia of Political Science. London: SAGE.

21.

Marchand

Morten

Shaw

(1999) The political economy of new regionalisms. Third World Quarterly 20 (5): 897–910.

22.

Marshall

Jaggers

Gurr

(2004) Polity IV. College Park, MD: University of Maryland.

23.

Mosley

(2003) Global Capital and National Governments. New York: Cambridge University Press.

24.

Pierson

(1996) The new politics of the welfare state. World Politics 48(2): 143–179.

25.

Pierson

(ed.) (2001) The New Politics of the Welfare State. New York: Oxford University Press.

26.

Rodrik

(1998) Why do more open economies have bigger governments? Journal of Political Economy 106(5): 997–1032.

27.

Rudra

(2002) Globalization and the decline of the welfare state in less-developed countries. International Organization 56(2): 411–445.

28.

United Nations Conference on Trade and Development (2014) UNCTADStat. Available at: http://unctadstat.unctad.org.

29.

Wibbels

(2006) Dependency revisited: International markets, business cycles, and social spending in the developing world. International Organization 60(2):433–468.

30.

Wibbels

Ahlquist

(2011a) Development, Trade, and Social Insurance. International Studies Quarterly 55 (1):125–149.

31.

Wibbels

Ahlquist

(2011b) Replication data for: “Trade, Development, and Social Insurance.” International Studies Quarterly 2011”, http://hdl.handle.net/1902.1/15466 V3 [Version].

32.

World Bank (2014) World Development Indicators. The World Bank. Available at http://data.worldbank.org/data-catalog/world-development-indicators.