What Determines the Depth and Institutional Strength of Regional Integration Agreements? A Joint International Economics / UNU-CRIS Seminar by Glenn Rayp

Abstract: Of the determinants of regional integration agreements (RIAs), Baier and Bergstrand (2004) argued that economic and geographical factors prevailed on political and institutional characteristics. As regards the likelihood of the formation of an agreement, this conclusion has been generally accepted in the literature. However, the proliferation in RIAs in the past decades reveals a very diverse set of institutional characteristics and levels of agreement scope and depth. This raises the question to what extent the depth and institutional strength of RIAs can be predominately explained by economic and geographical characteristics as well, i.e. from differences between the agreements in social welfare terms.

The evidence in this respect is fairly limited and, due to the lack of other data, merely considered the difference in the determinants of concluding deeper versus shallow agreements in terms of the Balassa hierarchy of regional integration. Márquez-Ramos et al. (2011) analysed the determinants of different stages in regional integration in a hierarchical order from free trade agreements (FTAs) to custom unions (CUs) and common markets (CMs). In each stage of integration, they found that socio-political and institutional factors had an impact on the likelihood of the agreement but were less important than economic and geographical characteristics. Endoh (2006) estimated separately the likelihood of the formation of deeper forms of integration such as FTAs and CUs compared to more shallow agreements based on the Enabling Clause. While confirming the overall impact of economic determinants, he found that political determinants only mattered for the first. However, taking into account their wildly differing levels within the distinct categories, the Balassa hierarchy may be a weak indicator of agreement depth and strength. The heterogeneity in terms of depth and institutional characteristics between agreements that are comparable in (Balassa) terms of economic integration, indicates that the determinants and their respective weights may be different between the formation and the depth and strength of an agreement.

In this contribution, we analyse empirically what may explain the depth and institutional strength of regional integration. Because we only observe the characteristics of integration between countries conditional upon the conclusion of an agreement, we estimate a two-step Heckman model of which the first step models the probability that two countries join a regional agreement and the second step the level of integration depth or institutional strength. As an indicator of integration depth, we use the Coredepth measure constructed by Hofmann et al. (2017) that counts the number of a maximum of 18 core provisions included in the agreement for all treaties currently operational and notified at the WTO. For the subset of regional organisations, i.e. the agreements between three countries or more providing an institutional superstructure, we derive an indicator of institutional strength derived from the CROP database (Jetschke et al., 2017, see also https : //lehrstuhlib.uni-goettingen.de/crop/information/ ), which is based on the information in the founding treaties of the organisations and major changes thereof. As explanatory variables, we include the set of economic, geographical, institutional and political country characteristics that are common to other studies in this field.

For both agreement depth and institutional strength, we find from several specification tests that political determinants have at least a comparable if not a larger impact than economic or geographical determinants and even these two combined. The log likelihood and the share of correctly selected second-step observations fall more and the average prediction error as well as the root mean-squared error increase more when we leave out the political variables instead of the economic or geographical determinants. The average level of democracy and governance quality has a significant positive impact on agreement depth, whereas the differences in democracy and governance between the partners have a negative impact. Rich and small countries also conclude deeper agreements, just as landlocked and more isolated countries. Distance has a negative impact as could be expected. As regards institutional strength, for the subset of regional organisations, all political determinants are found to have a negative impact. This would imply that countries with weaker institutions would tend to transfer more policy competences to regional organisations. Economic scale and differences in wealth have also a negative impact on organisational strength in addition to distance. In addition, contiguity, a landlocked location and sharing a common official language affect significant positively the institutional strength of regional organisations that countries. Our results are robust for different hypotheses regarding the second-step exclusion restrictions, in particular for agreement depth.

This is a closed event. For further information, please contact Ruben Dewitte (RubenL.Dewitte@UGent.be) and Pascale Vantorre in cc (pvantorre@cris.unu.edu).