Economic interdependence and international interactions: the impact of country size on political conflict and cooperation

Abstract

The trade–conflict model claims that one state, designated as the ‘actor’, is deterred from initiating conflict against a trading partner, designated as the ‘target’, for fear of losing the welfare gains associated with trade. This article extends the trade–conflict model to examine the effect of country size on the trade gains among countries. We derive three propositions with regard to international interactions that pertain to the links between trade, conflict and country size. These hypotheses all imply that a country with an improvement in its terms of trade with a large country will decrease conflict more than it would with an improvement in its terms of trade with a small country. A 30‐country sample from the Conflict and Peace Data Bank (COPDAB) is used for empirical tests. The empirical analyses support the derived hypotheses. The model predicts that a country's ability to influence domestic consumption in a trading partner is an important determinant of international interactions.

Notes

Detailed discussions and reviews are provided in Barbieri (1996), Barbieri and Schneider (1999), Reuveny and Kang (1998) and Mansfield and Pollins (2001).

For reference, see de Wilde (1991), Doyle (1997) and Spiegel (1991).

See Mansfield and Pollins (2001) and Polachek (forthcoming).

For example Rummel (1983), Chan (1984), Maoz and Abdolali (1989), Bremer (1993), Maoz and Russett (1993), Oneal et al (1996) and Polachek (1997) are a few of the many studies.

Reciprocity is the tendency of nations to play ‘tit for tat’, and inertia is the tendency of variables to depend on their own previous values (Reuveny 2003).

We can have m commodities, as in Polachek (1980). However, introducing such complications does not alter the analysis, merely the number of subscripts.

In international relations, it might be argued that if nations feel forced to choose conflict as the ‘least undesirable’ course of action available to them, there should always be some positive marginal benefit attached to this choice. One could further simplify by assuming that there is no direct welfare gain from conflict, ie W_z = 0. However, because we will define allies in terms of cross‐effects (W_zz ), we allow for W to be a function of Z (Polachek, Robst and Chang 1999). For the possibility that W_z < 0, see Polachek (1978, 1980).

Even if it is not a balance of payments, the derived results will not change.

We assume the marginal welfare gain from trade λ is constant across countries. While questionable, this assumption is necessary to reach the conclusions from the model. Also, we ignore the additional effect W_c (m′₁ − x′₁) on the welfare function.

The denominator is |H ₂| > 0 and the numerator is positive since λ > 0, x′₁ < 0, and [W_z2z2 + λ(x″ ₂ p _{x 2} − m″ ₂ p _{m 2} )] < 0 (Polachek, Robst and Chang 1999).

See Ray (1977) and Ethier and Ray (1979) for a different discussion of how gains from trade are related to country size.

It is worth noting that different assumptions can also make the proposition hold, but these complicated discussions are avoided here.

For a clear understanding of COPDAB, see Azar (1980) and Polachek (1980).

For the 30‐country sample, see Polachek (1997, 303, table 4).

A dyad with a high frequency of both conflict and cooperation still shows a moderate relationship, but only in the case of a high frequency of interactions. For further discussion, see Mansfield and Pollins (2001) and Polachek (forthcoming).

The events data, such as COPDAB, have limitations. For example, they are noisy, episode related, and not compatible (eg COPDAB and WEIS). However, the MID data set also has a couple of limitations including trade and conflict dynamics, conflict/cooperation interactions, the role of cooperation, and the relationship between trade and the levels of conflict and cooperation (as opposed to the probability of a MID). Such issues can be more readily investigated from events data. The two approaches need to be viewed as complements rather than substitutes. The events data collected in, or close to, real time have proved to be useful for both policy makers and academics. For a very detailed discussion, see Reuveny (2003).

For a detailed test, see Polachek (1997, 300).

It is worth noting that the empirical tests apply to average results. This research does not discuss the potential heterogeneity between nations that will cause different international interactions.

Yuan‐Ching Chang is an Associate Professor in the Economics Department at the Chinese Culture University in Taipei, Taiwan. His research interests encompass international interactions and economic interdependence. He is grateful to Solomon Polachek, John Robst and Peter Morton for their valuable comments.