742 Power or Plenty? not every state possesses.For example,the need to ensure continuing access to the region's oil resources might help explain current American security commit- ments in the Persian Gulf region. Historians of American foreign relations who emphasize the role of economic interests also suggest that the protection of important markets has sometimes led to American security commitments.For example,Hogan (1987,192)argues that the formation of the North Atlantic Treaty Organization was closely linked to efforts to bring about economic recovery in Western Europe and to integrate it into the international economic order American policy makers favored.Similarly, McGlothlen (1993)and Rotter (1987)link American security commitments in Asia to the effort to promote the recovery and integration of Japan.If arguments like these are correct,then part of the apparent effect of alliances on decisions about intervention is actually due to economic interests because these prompted the security commitment in the first place. The potential indirect effect of alliances through trade,and of trade through alliances,makes the magnitude and direction of the trade-alliance relationship important for assessing the effects of economic and security concerns on inter- vention.A single-equation model treating intervention as a function of all these variables together will not capture the indirect effects.A model that considers the potential indirect effects of these variables also suggests a different type of answer to the question of whether economic or security concerns are more important.Instead of emphasizing one or indicating that both influence policy makers in the same way,it holds out the possibility that the two considerations influence different stages of the policy-making process.The next section will present a research design for estimating these effects. Testing the Influence of Economic and Security Interests This section outlines a strategy for testing the role of economic and security con- cerns in various stages of the process leading to intervention.Figure 1 sets out the relationships that will be considered.The first stage of the test is to evaluate the direct effects of economic and security concerns on intervention decisions under certain circumstances.The causal paths of interest in this stage correspond to the arrows pointing to"U.S.Intervention'in the Figure.For the reasons discussed in the last section,concluding the analysis at this point,as previous quantitative analy- ses have done,neglects the important possibility that trade may influence interven- tion by shaping alliances,or vice versa.In order to capture this effect,the next step is to sort out the effects of trade and alliances on one another.Once this has been done,it is possible to estimate the total effect of the independent variables on intervention.The end result is a more realistic test of the relative importance of Economic Interests Rival Intervention U.S.Intervention Alliance commitment FIG.1.Influences on Intervention Decisionsnot every state possesses. For example, the need to ensure continuing access to the region’s oil resources might help explain current American security commit￾ments in the Persian Gulf region. Historians of American foreign relations who emphasize the role of economic interests also suggest that the protection of important markets has sometimes led to American security commitments. For example, Hogan (1987, 192) argues that the formation of the North Atlantic Treaty Organization was closely linked to efforts to bring about economic recovery in Western Europe and to integrate it into the international economic order American policy makers favored. Similarly, McGlothlen (1993) and Rotter (1987) link American security commitments in Asia to the effort to promote the recovery and integration of Japan. If arguments like these are correct, then part of the apparent effect of alliances on decisions about intervention is actually due to economic interests because these prompted the security commitment in the first place. The potential indirect effect of alliances through trade, and of trade through alliances, makes the magnitude and direction of the trade-alliance relationship important for assessing the effects of economic and security concerns on inter￾vention. A single-equation model treating intervention as a function of all these variables together will not capture the indirect effects. A model that considers the potential indirect effects of these variables also suggests a different type of answer to the question of whether economic or security concerns are more important. Instead of emphasizing one or indicating that both influence policy makers in the same way, it holds out the possibility that the two considerations influence different stages of the policy-making process. The next section will present a research design for estimating these effects. Testing the Influence of Economic and Security Interests This section outlines a strategy for testing the role of economic and security con￾cerns in various stages of the process leading to intervention. Figure 1 sets out the relationships that will be considered. The first stage of the test is to evaluate the direct effects of economic and security concerns on intervention decisions under certain circumstances. The causal paths of interest in this stage correspond to the arrows pointing to ‘‘U.S. Intervention’’ in the Figure. For the reasons discussed in the last section, concluding the analysis at this point, as previous quantitative analy￾ses have done, neglects the important possibility that trade may influence interven￾tion by shaping alliances, or vice versa. In order to capture this effect, the next step is to sort out the effects of trade and alliances on one another. Once this has been done, it is possible to estimate the total effect of the independent variables on intervention. The end result is a more realistic test of the relative importance of Fig. 1. Influences on Intervention Decisions 742 Power or Plenty?