Compensating for limitations in domestic output performance? Member state delegation of policy competencies to regional international organizations

1. Introduction

Cooperation in regional international organizations (RIOs) ¹ allows member states to work toward and perhaps achieve policy goals that would not be feasible unilaterally. Thus, it is not surprising that institutionalized cooperation between three or more states on the basis of regional proximity is neither a recent, nor a European phenomenon. ² Regional cooperation took place in two waves: after the end of WWII and again after the end of the Cold War, ³ resulting in 76 RIOs. ⁴ Member states initially equipped many of these RIOs with policy competencies in only few policy areas (e.g. economy and trade as prominent choices), but used treaty changes, annexes, appendices, and protocols as means to broaden the policy competencies over the course of time.

This article investigates one potential incentive for the delegation of policy competencies from member states to RIOs, namely whether states broaden policy mandates of RIOs in order to compensate limited domestic output performance. The comparative regionalism literature makes the argument that states that underperform with respect to the domestic provision of basic services have incentives to join RIOs and delegate policy competencies to them in order to compensate for their limited domestic output legitimacy. ⁵ This compensation can take different forms: first, collectively states might achieve policy results that they cannot realize on their own; second, states might receive support through their RIOs that helps to boost their domestic output performance, for instance through horizontal learning, best practices, or funds; third, states might get external recognition from third states and boost their domestic legitimacy in this manner; fourth, states might delegate policy competencies to RIOs in order to be able to engage in blame-shifting afterward – should RIO policies or activities meet domestic opposition; and fifth, states might delegate policy competencies to RIOs – even if these do not act on this basis – in order to showcase activity vis-à-vis their domestic constituencies and thereby window-dress their limited domestic performance. ⁶ Accordingly, the research question is, ‘Do states delegate policy competencies to RIOs in order to compensate for shortcoming in their own domestic output performance?’

This question is approached in four steps: first, the argument is explicated and potential alternative explanations are discussed in section ‘Theoretical accounts of states’ incentives to delegate policy competencies to RIOs ’. The subsequent section draws on a novel and comprehensive database, the Regional Organization Competencies (ROCO) database. It introduces the dependent variable of this article, member state delegation of policy mandates to RIOs, and presents the patterns of how states differ concerning how extensively they equip their RIOs with policy competencies in section ‘Equipping RIOs with policy competencies’. The hypotheses on state incentives to broaden policy mandates of their respective RIOs are subjected to an empirical plausibility probe in the subsequent section ‘Analysis and discussion ’. This reveals that incentives of states to furnish their respective RIOs with policy competencies vary. Concerning energy policy, states are generally the more willing to equip their RIOs with competencies the more domestic difficulties concerning output performance they face. Yet, in other policy fields, such as ‘development’ or ‘economy’, countries are more likely to cover corresponding policy competencies through their RIO, the better their domestic performance already is. This suggests that equipping RIOs with policy competencies is not a window-dressing exercise per se, used by states to compensate limitations of domestic output performance. The analysis also shows that democratic states are generally more likely to cover higher numbers of different policy competencies in their RIOs than autocratic states. This effect is even bigger the less output performance democracies achieve in agricultural or energy policies. Yet, in the majority of policy areas, democracies do not delegate competencies to RIOs to balance difficulties in domestic policy performance, but rather to further enhance the latter. A similar conclusion can be drawn for autocratic states. They equip their RIOs with more competencies when their output performance is limited in the environmental, energy, and finance realm. In all other policy areas, autocracies tend to delegate more competencies where their domestic performance is already better.

2. Theoretical accounts of states’ incentives to delegate policy competencies to RIOs

This article is x-centered in nature; ⁷ its research question seeks to uncover whether states delegate policy competencies to RIOs in order to compensate for problems in the domestic provision of public goods. RIO policy competencies are laid down in their primary law (e.g. Treaties, cf. section ‘Analysis and discussion’), which is created and changed in Conferences of the Parties (CoPs). In CoPs, every RIO member state has one vote and decisions are taken unanimously. Hence, RIOs can only be equipped with (additional) policy competencies, if each of the member states agrees to do so. Thus, the core hypotheses focus on conditions under which a state might be inclined to equip RIOs with policy competencies. ⁸ In a second step, it draws on state-of-the-art approaches on regional cooperation in order to identify control factors against which the core hypothesis will be empirically examined in section ‘Equipping RIOs with policy competencies’.

States can generate compliance with their laws on the basis of force, self-interest, or legitimacy. ⁹ While self-interest is volatile since not every citizen might consider each law as being in her own interest, and since governance by force would require unlimited enforcement resources and capacities, legitimacy is usually regarded as the most important driving force for state stability in the long run. ¹⁰ There is much literature on sources of legitimacy, defined as the generalized willingness of citizens to accept and comply with political decisions – irrespective of their content, ¹¹ many of which differentiate between input and output legitimacy. ¹² The former refers to the acceptance of the institutional design or the political system as appropriate in general, ¹³ and the ability of the citizens living in a political system to exert influence through democratic decision-making structures in particular. ¹⁴ Output legitimacy is the perception that a state performs well in generating and delivering public goods. ¹⁵ As data on output performance are available for a large number of states, over a long period of time, and for different policy areas, while this is not the case for output legitimacy (see section ‘Analysis and discussion’), the analytical and empirical focus of this article is on output performance, defined as the performance of states with respect to the provision of public goods.

Empirically, the performance of states varies when it comes to the provision of public goods, such as infrastructure, health care, or education, to name but a few. ¹⁶ Some states perform well across a broad array of policy areas and are likely to succeed to create output legitimacy on this basis. Others struggle to govern effectively in the domestic realm. ¹⁷ They have difficulties to provide public goods or might even be unable to do so at all, and are therefore unlikely to achieve high levels of output legitimacy.

Governments in states with low life expectancies, high infant mortality, high inflation, low levels of infrastructure investment, or low levels of trade might turn to RIOs in order to avoid electoral sanctions through their respective domestic constituencies – especially when elections are free, fair, and competitive. Delegating policy competencies to RIOs might be a means that allows states underperforming regarding the domestic provision of public goods to compensate for their limited domestic performance and thereby possibly also increase their output legitimacy. ¹⁸

There are several pathways of how states with limited output performance can benefit from delegating policy competencies to their RIOs. First, as classical approaches to international cooperation note, collectively states might be able to address problems and achieve results that they cannot realize on their own. ¹⁹ Hence, delegating policy competencies to RIOs in areas in which states do not perform too well on their own has the potential to increase their output performance – if the RIOs get active on the basis of their respective mandates. ²⁰ Second, whenever policies are costly in the short term, but might be beneficial in the long run, states might hesitate to enact them on their own, but instead delegate corresponding competencies to RIOs. ²¹ This allows for collective action and long-term solutions for underlying problems, whereas governments can blame the short-term costs arising from these policies on the RIO and avoid, thus, becoming subject to electoral sanctions through their domestic constituencies. ²² Such blame-shifting dynamics can, however, only take place if the RIO concerned does not remain inactive, but acts upon its policy mandates. Third, states with difficulties concerning a policy sector, might benefit from delegating policy competencies to their RIOs in this respect, since RIO policies might indirectly improve the member states’ domestic output legitimacy in the long run. ²³ For example, RIOs could serve as arenas for horizontal learning, best practices, or other new modes of governance. In addition, RIOs might even be useful for their member states to attract official development aid from third parties ²⁴ or might financially support their members themselves through structural funds for example. Fourth, even if RIOs remain inactive, states with limited output performance could still benefit from delegating policy competencies to them. As has been pointed out by scholars of regionalism, RIO membership and the equipment of RIOs with policy competencies can demonstrate the willingness of states to achieve policy goals or implement specific policies, ²⁵ and might boost the standing of member states in the international community in general, ²⁶ or with potential development aid donors in particular. Finally, states can have incentives to delegate policy competencies to their RIOs – even if these do not act on this basis – in order to showcase activity and their commitment to improve the domestic situation vis-à-vis their domestic constituencies. Yet, as long as the RIO member states do not collectively act on the basis of these RIO policy mandates later on, broadening the policy scopes of RIOs turns into a window-dressing exercise. These different incentives of states to equip their respective RIOs with policy competencies are not mutually exclusive, but very well compatible. Hence, hypothesis 1a states, ‘The more limited the policy performance of a state in a specific policy field, the more competencies it delegates to RIOs in this realm’ (H1a).

In any case, governments should be especially inclined to act on any or several of these potential incentives, the more democratic they are. Higher levels of democracy increase the chances that governments are sensitive toward possible domestic electoral sanctions ²⁷ and cover policy competencies in RIOs in order to further improve the provision of public goods or to compensate for shortcomings in the domestic realm. By contrast, governments in less democratic states might have less incentives to compensate limited domestic output performance and the associated output legitimacy by delegating policy competencies to their RIOs or to equip RIOs with additional policy competencies in order to further improve the provision of public goods. Accordingly, hypothesis 1b expects, ‘The more democratic a state is, the more policy competencies it delegates to RIOs’ (H1b).

Both components can also be combined through an interaction effect, leading to the third core expectation: ‘The positive effect of limited domestic policy performance on the delegation of policy competencies to RIOs is further accelerated, the more democratic the country in question is in nature’ (H1c).

The article controls for alternative explanations. First, sovereignty costs might differ, depending on whether RIOs are supranational or intergovernmental in character. In the latter, states remain in the driving seat and can control day-to-day RIO activities, while they at least partially give up authority in the former. The more supranational RIOs are in their day-to-day operation, the more hesitant states might be to equip them with additional competencies. In order to capture this, we control for supranational characteristics of RIOs (control 1). Second, power-based approaches stress that not all states might be equally benefiting from cooperation, as economically stronger states yield more power in the day-to-day operation of RIOs and, thus, over the policy outcomes. ²⁸ Thus, we control for economic power as well (control 2). Third, research on small states in world politics has demonstrated that on their own, small states can easily face difficulties when pursuing policy objectives on the international level. ²⁹ Thus, smaller states should benefit strongly from operating in and through RIOs rather than on a unilateral basis, ³⁰ and might delegate policy competencies to RIOs to this end. Hence, we also control for state size (control 3). Fourth, hegemonic states might try to use the RIO to pursue their self-interests and prevent unfavorable policies. Thus, we control for the number of hegemons in each RIO, in general, as well as whether or not Russia is member of a RIO, in particular, to capture the potential impact of autocratic hegemons (control 4).

Before the plausibility of the hypotheses can be empirically examined while controlling for the identified factors, the next section introduces the dependent variable: the policy competencies that states equipped their RIOs with.

3. Equipping RIOs with policy competencies

When RIOs are defined as institutionalized forms of cooperation between three or more states that are located in geographical proximity and have a headquarter location and often also a secretariat, there are a total of 76 organizations that meet these criteria (cf. Table 2). What RIOs are able to do in theory and what they actually do in practice differs widely. Yet, the policy mandates of RIOs usually form the basis of their policy-making and actions as they express a commitment of RIO member states toward common regional activities and form the legal basis for such activities. Policy mandates are laid down in RIOs primary law, which encompasses treaties, annexes, appendices, or protocols. The ROCO database is based on these primary law sources for all 76 RIOs and entails information on the policy competencies of all 76 RIOs between 1945 and 2015, other RIO characteristics, and state membership in RIOs over time.

This article sheds light on nine different policy fields: agriculture, development, economy, energy, environment, finance, good governance, health, and technology. These fields were selected as they cover core areas in RIOs all across the globe and vary concerning characteristics, such as being primarily regulative (e.g. technology, health) or having pronounced distributional effects (e.g. development, agriculture). For each policy area, we identified between 14 and 17 specific policy competencies ³¹ on the basis of a computer-assisted text analysis. Once the relevant articles or paragraphs were identified, we hand-coded the respective RIO primary law sources in order to ensure context-sensitive coding ³² and to detect which specific internal and which specific external policy competencies a RIO possesses in a given year and policy field. ³³ The dependent variable of this article is the total number of specific policy competencies in the fields of agriculture, development, economy, energy, environment, finance, good governance, health, and technology that a state covers in the RIOs of which it is a member in a given year (cf. Table 4, Appendix 1, with descriptive statistics).

Figure 1 illustrates that the number of policy competencies that states have equipped their respective RIOs with increases considerably over time. In 1945, the average state covered 0.41 policy competencies in its RIOs, which increases to 3.39 in 1955, to 10.61 in 1965, to 31.36 in 1975, to 44.51 in 1985, to 102.92 in 1995, to 163.53 in 2005, to an average of 204.80 policy competences in 2015. The number of policy competencies increases gradually until 1975 and rises strongly from 1985 and especially from 1995 onward. Thus, the first period of transferring competencies to RIOs starts after the end of WWII and is characterized by an incremental approach, while the number of policy competencies that states have covered in their RIOs increases sharply after the end of the Cold War.

OPEN IN VIEWER

Figure 1.

Sum of delegated policy competencies per state, nine policy fields, selected years.

Apart from this temporal variation, Figure 1 shows that some states have delegated considerably more policy competencies to RIOs than others. In 1945, the states with the highest number of policy competencies covered in RIOs are Iraq, Jordan, Lebanon, Saudi Arabia, Syria, and Egypt with 13 competencies each. In 1955, France was the country with the highest number of competencies covered in RIOs (45), followed by the Netherlands (41), the United States (29), and then Luxembourg, Italy, Germany, and Belgium (28 each). Ten years later, the Netherlands, Luxembourg, and Belgium each have delegated 83 policy competencies to RIOs, while France scores 68, followed by Nicaragua, Honduras, and Guatemala with 52 competencies each. By 1975, 13 countries have transferred more than 100 competencies to their RIOs. These are Sweden (109), the United States (110), Iceland and Norway (111 each), Ireland (113), Germany and Italy (115 each), France and the UK (132 each), Denmark (136), as well as Belgium, Luxembourg, and the Netherlands (147). By 1985, there are a total of 19 countries that have delegated more than 100 policy competencies to RIOs. In addition to the countries with high scores in 1975, in 1985, Bolivia, Columbia, Ecuador, Peru, and Venezuela have also covered more than 100 policy competencies in RIOs (103 each, to be precise). In 1995, the number of policy competencies delegated to RIOs increases sharply. Hundred countries have equipped their respective RIOs with 100 or more competencies, and 19 even reach scores of 200 and higher. These are Sweden and Finland (287 each), Denmark (270), the Netherlands, Luxembourg, and Belgium (257), Germany (249), France (245), Austria (242), the UK (238), Spain, Portugal, Italy and Greece (225 competencies each), Ireland (223), the United States (220), Canada (207), as well as Norway and Iceland (204 each). Ten years later, in 2005, 63 states have transferred 200 and more policy competencies to their respective RIOs, with nine countries exceeding the 300 mark. These are Russia (453), Denmark (348), Sweden and Finland (346 each), the United States (344), Kenya (334), Canada (331), Uganda (322), and Sudan (310). Finally, in the year 2015, in which the average state has delegated more than 200 policy competencies to RIOs, the highest scores are reached by Russia (651), followed by Kyrgyzstan and Kazakhstan (467 each), Tajikistan (443), and Uzbekistan (440).

There is considerable temporal variation with respect to how many and which policy competencies states cover in their RIOs at a given point in time. In the post WWII international system, the overall number was considerably lower than in the post-Cold War era. In the early period, states incrementally broadened RIO policy scopes; after 1990, a sharp increase in RIO policy competencies took place. In 1945, states had delegated a total of 81 competencies to RIOs, which increased over time to a total of 671 in 1955, to 2101 in 1965, to 6210 in 1975, to 8812 ten years later, to 20,379 in 1995, to 33,147 in 2005, and finally to a total of 40,552 policies that states cover in their respective RIOs in 2015. There is also variation between policy fields (Figure 2). In general, states have always been more inclined to transfer economic policy competencies to their RIOs, followed by competencies in the realm of technology. Environmental policy competencies – although making it to the third rank in the later period of regional cooperation – were not delegated to RIOs prior to 1974 at all, but increased particularly steeply in the early 1990s. Finance, good government, and the developmental policy area as well as energy form a mid-field, while states are less inclined to equip their RIOs with health and agricultural competencies.

OPEN IN VIEWER

Figure 2.

Aggregated member state delegation of policy competencies to RIOs over time.

Zooming in on the nine policy areas of interest in this article, there is also considerable in-country variation. Figure 3 illustrates this variation for eight different larger and smaller states that also vary with respect to their regime types in the year 2015. Russia, for instance, covers high numbers of policy competencies in RIOs in the fields of technology and economy, followed by the environment, while the RIOs Russia is a member of cover agricultural competencies considerably less often.In general, China has equipped its RIOs with fewer competencies than Russia. In particular, China has refrained from delegating good governance and environmental policy competencies to its RIOs. The United States especially covers economic and environmental competencies in RIOs, followed by technological competencies. Finance, good governance, and agriculture, however, are policy fields that RIOs, which the United States has joined, are not strongly equipped with. Brazil places much more emphasis on economic policy competencies than on good governance, finance, and technology. Denmark covers almost as many competencies in the environmental policy field in its RIOs as in the policy fields, economy and technology combined. Germany’s preference for delegating policy competencies to RIOs looks similar to the Russian preferences, although in general Germany covers fewer policies in RIOs than Russia. Uganda covers developmental and health competencies least often in the RIOs it has joined, preceded by competencies in the policy field of agriculture. Finally, South Africa also resembles the pattern of in-country variation of Russia and Germany, but on a lower scale and with two differences: energy and environmental policy competencies were delegated less often to RIOs – compared to economy and technology competencies.

OPEN IN VIEWER

Figure 3.

Radar plots, selected countries, 2015.

4. Analysis and discussion

This section first discusses the operationalization of the variables and the model selection. In a further step, the hypotheses are put to an empirical plausibility probe and the findings are discussed.

Working with cross-sectional time series data that cover all states between 1945 and 2015 is a challenge. Thus, when faced with choices concerning the operationalization of the independent variables, the indicator, for which the data coverage is best in terms of the timeframe and countries, was selected (cf. Table 4, Appendix 1, for descriptive statistics).

H1a focuses on output performance of states and expect that ill-performing states have stronger incentives to equip RIOs with policy competencies in the corresponding policy area, either in order to improve the provision of public goods through RIO policies or activities directly or indirectly through horizontal learning, to engage in blame-shifting, or in window-dressing behavior. The core independent variable of the hypothesis is a country’s output performance. This is measured for each of the nine policy areas separately. The selected performance indicators are available over a long period of time and capture an important aspect of state performance in the policy area in question. In the realm of agriculture, the proxy used for output performance is the income from agriculture relative to the size of arable land in a country (in million USD/a country’s arable land in km²). The data, based on which the variable agricultural income in USD was created, stem from the World Bank. The components gross domestic product (GDP) (in million USD see below) and share of agriculture as percentage of GDP are multiplied in order to obtain information on the USD that a country’s agricultural sector produces in a given year and have been obtained from the World Development Indicators of the World Bank. This term is divided by the arable land in square kilometers of a country, which in itself is calculated by multiplying the size of a country by the share of arable land (data for these components have been obtained from the World Bank). The proxy capturing performance in the development policy field is life expectancy in years. The data also stem from the World Bank’s world development indicators series. In the policy area of economy, performance is captured by the extent of trade per country and year (as percentage of GDP) and also stems from the World Bank’s world development indicators series. This is the same source as for the independent variable energy import measured in percent of energy consumption, which captures the extent to which countries are not self-sufficient, but dependent on imports. The proxy ‘deaths caused by environmental disasters’ captures the total deaths from meteorological crises in a given year and stems from the United Nations Statistics Division (UNSD). It is a negative proxy for the policy performance of a country in the realm of environmental policies. A country’s performance in the policy area ‘finance’ is captured by the inflation of consumer prices in percentages. The data stem from the World Bank’s world development indicators series as well. The variable ‘executive bribes’ serves as a measure for good governance performance and is taken from the varieties of democracy project. ³⁴ In the health policy area, the indicator selected for the measurement of performance is infant mortality. It captures the number of infant deaths per 1.000 births. Again, the data stem from the World Bank’s world development indicators series. In the policy area ‘technology’, the indicator selected to capture performance is the expenditure for education as percent of GDP, which has been obtained from the World Bank’s world development indicators series.

The independent variable of H1b is the regime type of countries. This is measured by the extent to which a state is autocratic or democratic, the data for which stem from the Polity IV project. ³⁵

The control variables are measured as follows: the supranational character of RIOs is operationalized in two ways, first capturing whether majority decisions are possible in each RIO and year, and second, measuring whether or not a regional court is present. The data of both dummy variables stem from the ROCO database. The second control variable, the economic strength of countries, is measured by a country’s GDP in million USD at constant 2005 national prices. The data stem from the Quality of Government OECD dataset 2016. ³⁶ Country size is captured by the number of inhabitants in billion in a given year. The data stem from the World Development Indicators dataset of the World Bank. Finally, two indicators control for hegemony, one counting the number of hegemons in the RIOs a state is member of (China, Russia, United States), and second counting the number of RIOs a state has joined in which Russia is also a member state. The data stem from the ROCO database.

The article’s dependent variable is the number of different policy competencies that a state covers in its RIOs at a given point in time in each of the nine different policy areas, respectively (see section ‘Equipping RIOs with policy competencies’). Primary law constitutes the rules of the game and the obstacles of changing them are very high, requiring unanimity among all RIO member states. Thus, if even a single member state disagrees to changing primary law in CoPs, RIOs cannot be equipped with additional policy competencies. Since each RIO member state has to agree to each change of RIO policy competencies, this article uses the individual state agreement to equip an RIO with policy competencies as the unit of analysis. The dependent variable is of count character and is overdispersed. Accordingly, negative binomial time series cross-sectional regression models are fitted in which the independent variables are lagged by 2 years (the findings remain robust, if the independent variable data are lagged 5 years, cf. Table 6, Appendix 1) and standard errors are grouped on the state-level, as there are repeated observations for each state in the dataset.

H1a expected that states are more likely to equip their RIOs with additional policy competencies, the less well they perform domestically. Yet, as Table 1 shows, this is not the case in all policy areas. Only in the policy field ‘energy’ (Model 4, Table 1), negative incentives might be at play as the finding is highly significant and the sign points to the expected direction. Higher levels of being dependent on the import of energy significantly increases the propensity that the respective states equip their RIOs with energy-related policy competencies. In this policy area, limited output performance might operate as incentive to delegate competencies to RIOs, since a regional energy policy can be more beneficial for the member states than acting alone, for instance, when building the infrastructure for energy imports or when negotiating energy prices with third-party exporters.

OPEN IN VIEWER

Table 1.

Regressions. ^a

	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7	Model 8	Model 9
	Agriculture	Development	Economy	Energy	Environment	Finance	Good governance	Health	Technology
Agriculture income	0.000***
	(0.000)
Life expectancy		0.062***
		(0.002)
Trade			0.003***
			(0.000)
Energy imports				0.000***
				(0.000)
Environmental deaths					−0.000***
					(0.000)
Inflation						0.000
						(0.000)
Executive bribes							−0.127***
							(0.015)
Infant mortality								−0.012***
								(0.000)
Education expenditures									0.036***
									(0.005)
Democracy	0.028***	0.026***	0.028***	0.048***	0.057***	0.027***	0.046***	0.023***	0.033***
	(0.002)	(0.001)	(0.002)	(0.003)	(0.003)	(0.002)	(0.002)	(0.002)	(0.002)
RIO supranationality	0.414***	0.203***	0.379***	0.295***	0.207***	0.499***	0.487***	0.142***	0.336***
	(0.010)	(0.008)	(0.007)	(0.014)	(0.013)	(0.008)	(0.009)	(0.008)	(0.011)
Hegemons in RIOs	0.217***	0.211***	0.130***	0.293***	0.229***	0.212***	0.198***	0.358***	0.209***
	(0.009)	(0.004)	(0.004)	(0.007)	(0.007)	(0.005)	(0.006)	(0.006)	(0.006)
Population	0.000***	0.000***	0.000***	0.000***	0.000***	0.000***	0.000***	0.000***	0.000***
	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)	(0.000)
Constant	1.141***	−1.463***	0.490***	−0.221***	0.433***	0.303***	0.401***	4.964***	1.011***
	(0.067)	(0.100)	(0.031)	(0.048)	(0.047)	(0.039)	(0.046)	(0.397)	(0.055)
Observations	4823	7525	6594	5113	3708	6128	7353	7184	2855
BIC	22,283.104	36,853.833	49,282.760	24,865.488	24,457.134	35,865.346	36,658.087	30,300.874	19,578.842
LL	−11,111.868	−18,391.212	−24,610.601	−12,398.586	−12,195.694	−17,902.151	−18,293.432	−15,114.919	−9757.594

RIO: regional international organization; BIC: Bayesian information criterion; LL: log-likelihood

Standard errors in parentheses with * p < 0.05, ** p < 0.01, *** p < 0.001.

a

The control variables, country size and economic power, correlate highly as do the two measurements of RIO supranationality (majority decision-making rule; presence of a regional court). Thus, these pairs of variables are put in different models in order to avoid problems arising from multicollinearity. As a robustness check, the regressions were also run with the economic power variable and the measurement of RIO supranationality through regional courts and a time lag of 2 years (cf. Table 5, Appendix 1) as well as of 5 years (Table 6, Appendix 1). All findings remain robust and the significance levels do not change either.

Yet, there is no empirical evidence that limited output performance triggers policy scope expansion in the areas of agriculture, development, economy, environment, finance, good governance, health, and technology. In fact, some signs point to the direction opposite of our expectations. For example, in the policy area ‘agriculture’, states have a higher propensity to cover high numbers of agricultural competencies in their RIOs the higher their agricultural income (Model 1, Table 1). This might suggest that states with ill-performing agricultural sectors do not expect much gains from regional cooperation per se, that the saliency of agricultural policies is too limited in such countries to trigger state compensation activities on the RIO level, or that democracies and autocracies behave differently (see discussion below).

In the field ‘development’, states facing fewer problems are significantly more likely to delegate higher numbers of developmental competencies to their RIOs (Model 2, Table 1). The same applies to countries with high levels of economic performance (Model 3, Table 1). Both findings are counterintuitive, as expected benefits from cooperation in RIOs should be high in the developmental and the economic realm, not in the least since many RIOs started in these two policy areas. ³⁷ Instead of compensation mechanisms being at play, the findings suggest that there is a positive spiral in these policy areas: the better off states already are, the more cooperation in RIOs they pursue.

In the policy field environment, the lack of domestic performance decreases rather than increases the propensity of states to equip their RIOs with many environmental policy competencies (Model 5, Table 1). This could be reflective of the fact that environmental policies on the RIO level can become costly. ³⁸ Especially states with limited environmental performance might face high costs when environmental standards are strengthened and might therefore be less likely to cover such policy competencies in their respective RIOs in the first place.

Executive bribes significantly reduce the likelihood that states equip their RIOs with high numbers of good governance competencies, which is also not in line with H1a (Model 7, Table 1). Again, this finding might be due to potential costs that governments in countries with high levels of corruption associate with equipping RIOs with good governance competencies. ³⁹ In such countries, governmental actors might have more to lose than to gain from equipping RIOs with policy competencies, that at the very least signals the inappropriateness of bribing to their respective constituencies, and eventually might even lead to RIO policies or activities geared toward combatting corruption and promoting good governance.

H1a would also have expected that decreasing levels of domestic output performance in the finance, health, and technology policy fields increase the probability of states equipping their RIOs with such competencies, respectively. Yet, there is no empirical evidence pointing toward the plausibility of the compensation mechanism in these policy areas (Models 6, 8, and 9, Table 1). In fact, as models 8 and 9 (Table 1) show, the opposite is the case. Countries are more inclined to cover health and technology competencies in their RIOs, the better they already perform in these respects. Thus, instead of a compensation mechanism, a positive spiral seems to be at play in the health and the technology policy fields as well. In the finance realm, the sign points to the expected direction, but the covariate is not significant in any of the model specifications (Table 1 and Appendix 1 Tables 5 and 6). ⁴⁰

In all models, an increase in democracy robustly and significantly increases the likelihood that states cover an increasing number of policy competencies in their respective RIOs. This is in line with H1b. Thus, the more democratic states are, the higher their propensity to equip the RIOs they are members of with agricultural, development, economic, energy, environmental, financial, good governance, health, and technology competencies.

The findings of hypotheses 1a and 1b cannot be interpreted as strongly supporting the notion that states equip their RIOs with policy competencies across the board, in order to avoid electoral ex post sanctions by compensating limited domestic output performance. This is likely to be the case in one policy area (energy), while for the bulk of policy fields (agriculture, development, economy, environment, finance, good governance, health, and technology), the analysis suggests that democracies tend to further increase the well-functioning provision of public goods by equipping their respective RIOs with corresponding policy competencies.

Based on Table 1, limited domestic output performance alone does not seem to be a major driving force behind regional integration in all policy fields, as only two of nine policy areas would be in line with this expectation. Yet, maybe domestic shortcomings in output performance only matter if governments are subject to free, fair, open, and competitive elections; in other words, if states are democratic in nature, they are more sensitive toward potential losses in output legitimacy. In order to shed light on this potential interaction (H1c), we opted for a graphical analysis of the interaction effects between regime type and output performance for each of the nine policy areas separately. Figure 4 shows the predictive margins between the regime type and the output performance. Thereby, we distinguish between highly democratic (+10), partially democratic (+5), partially autocratic (−5), and highly autocratic countries (−10) and plot their respective effects moderated by output performance in each of the nine policy areas on the dependent variable, the number of policy competencies with which states equipped their RIOs in the policy area at stake.

OPEN IN VIEWER

Figure 4.

Predictive margins.

Figure 4 illustrates that it is not the case across all policy areas that the two core independent variables, the output performance and the regime type of countries, interact in the expected manner: namely that the positive effect of decreasing output performance on the likelihood of states delegating corresponding policy competencies to their RIOs is further accelerated the more democratic a state is. In the policy areas, development, economic, environment, finance, good governance, health, and technology, H1c needs to be rejected. It is not the case that under conditions of low domestic performance (i.e. life expectancy below 40 years), democracies are more inclined to compensate their limited output performance through delegating development competencies to their RIOs.

With respect to the economic policy field, H1c is not supported by empirical evidence. Countries with lower shares of trade as percentage of GDP (less than 50) are increasingly more inclined to delegate economic competencies to RIOs, the more democratic they are. However, even for countries with higher scores of trade, democracies tend to be more inclined to cover economic competencies in RIOs than autocracies. Thus, it is not the case that the likelihood of equipping RIOs with competencies decreases especially for democracies as the domestic economic performance increases. Instead of a compensation mechanism being prevalent, Figure 4 indicates a positive spiral according to which all states are generally more inclined to equip RIOs with economic policy competencies, the better their domestic performance becomes.

Figure 4 shows that H1c also needs to be rejected with respect to the policy field environment. In fact, an opposite interaction effect can be evidenced empirically. The lower a country’s environmental performance, the less inclined democracies become to delegate environmental competencies to their respective RIOs, while autocracies become more willing to cover environmental policies in their RIOs. This counterintuitive finding could indicate that there are differences between ‘democratic RIOs’ and ‘authoritarian RIOs’. ⁴¹ RIOs composed of democratic states are more likely to get active on the basis of their competencies than ‘autocratic RIOs’ and are also more likely to pass binding highly regulated and, thus, costly rules. Hence, democracies with limited domestic performance shy away from delegating environmental competencies to RIOs, whereas autocracies with limited domestic output performance might encounter less costs when equipping their RIOs with environmental competencies, but could gain legitimacy or reputation-related benefits.

In the realm of finance, the positive effect of low domestic performance on a states’ propensity to equip their RIOs with financial competencies is moderated by regime type. The higher inflation becomes, the higher the propensity of autocracies to delegate financial competencies to RIOs. This interaction between domestic performance and regime type is not in line with H1c, as democracies do not engage in more compensation dynamics than autocracies the poorer their domestic output performance becomes.

Democracies delegate more good governance competencies to their respective RIOs than autocracies. Yet, as levels of corruption increase, chances that states equip their RIOs with good governance competencies decline. Although this decrease is steeper the more autocratic countries are, it still does not provide empirical support for H1c, which would have expected a rise in delegation of competencies to RIOs as domestic performance declines.

The same conclusion can be reached for the policy area health. The higher the infant mortality and the worse the policy output in this respect, the lower the number of health competencies that democracies are covering in RIOs, while autocracies with equally low levels of output performance are much more inclined to equip their RIOs with health competencies.

The policy area education does not exhibit an interaction effect corresponding to the expectation of H1c. Irrespective of education expenditures, democracies have higher probabilities to transfer educational policy competencies to RIOs than autocracies.

In two policy areas, namely, agriculture and energy, there are interaction effects in line with H1c. In the policy field ‘agriculture’, the expected interaction effect is apparent for very low levels of agricultural performance. If the income from a country’s arable land is lower than US$10 million, the positive effect of limited domestic policy performance has on a country’s delegation of policy competencies to its RIOs is furthered, the more democratic the country in question is. Thus, in the policy field of agriculture, it is plausible that states with low levels of output performance are especially inclined to compensate shortages by covering this policy area in RIOs, possibly in order to avoid electoral sanctions directly through RIO performance, or indirectly through blame-shifting. For levels of agricultural income that exceed about US$10 million, it is, however, not the case that democracies are more inclined to transfer agricultural competencies to RIOs than autocracies. In fact, under conditions of higher agricultural outputs, autocratic regimes are more willing to cover agriculture competencies in their respective RIOs than democracies.

Figure 4 illustrates that with respect to energy policies, democracies are generally more inclined to cover this policy area in the RIOs they belong to than autocracies. The more energy a country needs to import, the higher the likelihood that it equips its RIOs with energy policy competencies, and this is further increased the more democratic the states in question are. Zooming in on the right part of this graph, in which countries depend more heavily on energy imports, reveals that regime type distinctions get smaller, but that democracies are still more likely to delegate energy competencies to their RIOs than less democratic regimes.

In addition, Figure 4 reveals that in the majority of policy fields, namely development, economic, good governance, health, and technology policies, both democracies and autocracies do not increasingly delegate competencies to RIOs in order to balance difficulties in domestic output performance. Instead, they broaden RIOs’ policy scopes more strongly the higher the domestic performance already is. Finally, autocracies are likely to engage in compensating efforts in the environmental, the energy, and the finance policy area. The worse the domestic output in these three policy fields, the higher the propensity of autocracies to equip their RIOs with corresponding policy competencies.

5. Conclusion

Comparative regionalism constitutes a research agenda that has attracted increased attention over the last decade and led toward the comparative analysis of RIOs. This article adds to this strand of scholarship as it examines the development of policy competencies of 76 RIOs over an extended period of time (1945–2015). This reveals that, over time, the scope of RIO policy competencies gets broadened as member states tend to increasingly delegate competencies to their RIOs. Despite this temporal trend, there are distinct patterns. First, some RIOs possess more policy competencies than others. Second, not all states are equally inclined to cover equally many policy competencies through their respective RIOs. Third, not all policy areas are equally often delegated to RIOs. Based on a state-centered perspective on regional cooperation, this article inquired into one specific potential driving force behind varying states’ inclination to equip RIOs with policy competencies in different policy fields. Accordingly, the question on the center stage is as follows: ‘Do states delegate policy competencies to RIOs as a means to compensate for limited domestic output performance?’.

In order to answer this x-centered research question and shed light on this potential motive behind the broadening of RIO policy scopes, the article first explicated the core hypotheses. According to these expectations, the propensity of states to equip their RIOs with policy competencies increases, the lower the domestic output performance in a policy field, the more democratic countries are, so that governments seek to avoid electoral sanctions, and if an interaction between regime type and domestic output performance takes place. These hypotheses were subsequently put to an empirical test.

This revealed that only in one policy area, namely energy, are states generally more willing to equip their RIOs with corresponding competencies, when they face domestic difficulties concerning output performance. By contrast, in the policy areas of development, economy, environment, good governance, health, and technology no such effect can be evidenced for all states. Instead, the better the domestic output in any of these policy fields already is, the more likely states are to equip their RIOs with higher numbers of competencies in the corresponding policy field. Distinguishing further between regime types reveals that in the majority of policy areas, namely in development, economic, good governance, health, and technology policies, neither democracies nor autocracies delegate competencies to RIOs in order to balance difficulties in domestic policy performance, but rather to further enhance the latter. Autocracies are more likely to equip RIOs with environmental, energy, and finance policies, the worse their domestic output performance in these policy areas is. Similarly, democracies are more likely to equip their RIOs with competencies, when their domestic output performance is decreasing in agriculture and energy policies.

This analysis suggests that in the majority of policy areas, namely in development, economic, good governance, health, and technology policies, neither democracies nor autocracies delegate competencies to RIOs in order to balance difficulties in domestic policy performance, but rather to further enhance the latter.

Some scholars have pointed to the possibility that states might equip RIOs with policy competencies in order to compensate limited domestic output performance – either directly through RIO policies or other RIO activities, through being able to engage in blame-shifting to RIOs should short- or medium-term costs arise from RIO policies, through attracting external funding, through blaming RIOs for a standstill in policy-making, or through showcasing activity vis-à-vis domestic audiences and thereby avoid electoral sanctions (window-dressing). ⁴² The latter three aspects might at least partially explain a seemingly counterintuitive finding, namely that RIOs differ considerably concerning whether they indeed act on the basis of their primary laws. ⁴³ It might be indicative of window-dressing, if especially states with limited provisions of public goods that are faced with potential electoral sanctions or sanctions from third parties (e.g. donors) broaden the policy scopes of their RIOs. Vice versa, delegating competencies to RIOs under conditions of high domestic outputs or in the absence of potential sanctions do not suggest that window-dressing is taking place.

There are policy field specificities at play. Most importantly, democracies in the agriculture and energy fields and autocracies in the environmental, energy, and finance fields are more inclined to equip their RIOs with additional competencies, the worse their domestic output performance in these respective policy areas. In these instances, it is possible that governments seek to compensate potential shortages of domestic output performance through equipping RIOs with the means to get active in these policy fields and avoid potential sanctions from the electorate or third parties in this respect. This allows to shift the blame for lacking output performance at least partially to RIOs – should RIOs remain inactive, be unable to solve the problem, or solve the problem but do so by conferring costs on the member states and their citizens. However, window-dressing is not the only possible interpretation. The inclination to equip RIOs with competencies when democratic governments perform badly in agriculture and energy fields or when autocratic governments struggle in the environmental field could also stem from the belief and desire of states to collectively remedy the problems underneath limited domestic output performance in these fields, either through RIO policies or through RIO actions. Uncovering which of these rationales ultimately guides these governments in these policy areas calls for in-depth qualitative case studies, which are beyond the scope of this article and up for future work.

In any case, in the majority of policy areas, namely in development, economic, good governance, health, and technology policies, neither democracies nor autocracies increasingly delegate competencies to RIOs in order to balance difficulties in domestic policy performance, but rather to further enhance already good domestic performance. Instead, they are more likely to extend RIO policy scopes, the better the domestic performance already is. This suggests that more often than not, the broadening of policy scopes is not a symbolic exercise, but reflects the intention of states to indeed use RIOs as a means to achieve positive contributions to the domestic provision of public goods. ⁴⁴ Also, states with high levels of domestic output performance in a policy area might equip their RIOs with many corresponding policy competencies in order to operate as policy entrepreneur. This allows standard-setting and ensuring that other RIOs member states face the same potential costs associated with high standard regulations in policy areas such as health, environmental, or good governance policies. ⁴⁵

The findings carry an additional important implication. Autocratic states tend to perform less in the provision of public goods than their democratic counterparts, ⁴⁶ which goes hand in hand with more limited policy scopes of RIOs and, thus, fewer opportunities to improve policy performances through concerted activities on the RIO level. This can contribute to a vicious cycle: The worse countries are off, the less likely they are to use RIOs as a means to work for collective solutions in equipping them with policy competencies – although RIOs can be highly beneficial to their member states. ⁴⁷