Note: This is quite an old summary. I wrote this when I was a freshman (2016), so it doesn’t read as smoothly as my current writing.
Also, one of the constraints was for the summary to be written in a single page. Hence, the horrible formatting.
McDonald’s (2004) research question asks, “Is free trade, and not just trade, the key to dyadic peace?”. His motivation for the study stems from a lack of consensus concerning the precise nature of the link between commerce and conflict. Consequently, McDonald argues that free trade promotes dyadic peace by removing protective barriers. To clarify, he argues that protective barriers play an important role in decisions to use military force, e.g., enhance the domestic power of societal groups likely to support war, reduces the capacity of free-trading interests to limit aggression in foreign policy, and simultaneously generates political support for the state often used to build its war machine. Furthermore, McDonald argues that his mechanism connecting free trade to peace is more relevant than others because it includes state-society interactions to explain foreign policy behavior, which others have failed to include. By including state-society interactions, McDonald is able to explain how trade promotes peace by connecting the domestic distributional consequence of commercial policy with the domestic politics of international conflict.
McDonald’s hypothesis is that greater levels of protection increase the probability of interstate conflict. Given the dichotomous dependent variables, McDonald estimated his model with logistic regression and the Beck, Katz, and Tucker (1998) correction for time- series, cross-sectional analysis with a binary dependent variable. His unit of analysis is the dyad-year while his temporal domain is a function of how protection is measured, e.g., when using data on import duties, the cases span from 1970 to 2000. However, data utilizing the Hiscox and Kastner indicator (gravity model of trade) are available from 1960 to 1992. Moreover, all interstate dyads in the international system provide the spatial domain of cases. The dependent variable is the onset of a new militarized interstate dispute between dyad members i and j, sourced from the revised MID data version 3.0 and derived from Eugene 3.03 (Bennett and Stam 2000). As a categorical variable, MIDON, takes on a value of 1 in the first year of a new dispute. McDonald’s main independent variable is protection or openness, which is operationalized in two ways. The first indicator of protection (PROTECTH), with data obtained from the World Bank Development Indicators, measures the proportion of customs revenue divided by total imports in the state possessing the greater such ratio in the dyad. The second indicator (for data from Hiscox and Kastner, 2002) follows a similar logic; they present protection scores as deviations from the sample maximum, or free trade, state in their data set – the Netherlands in 1964. Both indicators of protection (PROTECTH) are continuous variables and measure the score of the state in the dyad that possesses higher barriers to trade. Additionally, PROTECTH is measured in period t – 1 to account for potential endogeneity effects between conflict and protection. To ensure the robustness his results, in the baseline regression model, McDonald controls for the level of democracy (DEMOCRACYL), relative size of the largest economy (GDPH), changes in short-term economic conditions (GROWTHL), similarity of political interests (ALLY), alliance portfolio similarity (INTERESTS), differences in capabilities (CAPRATIO), presence of a great power (MAJORPOWER), geographical contiguity by land (CONTIGUITY), and logged distance (DISTANCE).
McDonald’s results show that PROTECTH is consistently positive and statistically significant, indicating that the probability of a new military dispute increases as the level of protection increases. These results support the domestic version of commercial liberalism presented by McDonald. Additionally, he finds that when including both PROTECTH and DEPENDL in his model, the coefficient on DEPENDL shrinks and is no longer statistically significant. This casts doubts on the opportunity cost and sociological variants of commercial liberalism. Overall, his results seem to support the claim that free trade enhances the prospects for peace.
McDonald, P. J. (2004). Peace through trade or free trade?. Journal of conflict resolution, 48(4), 547-572.