鈴木 一敏

The U.S.-Japan Structural Impediments Initiative (SII) included six different areas. Although negotiated simultaneously, the distributive results of each area were quite different. Previous studies have attempted to explain this outcome by comparing each area's internal conditions (e.g., the degree of participation expansion in the Japanese political process, the similarity of the perceptions of two government officials regarding the problem involved). However, this research design is inappropriate for an omnibus negotiation like SII. For a comparison to be valid, the outcomes of each area have to be independent. If there is a cross-area barter, the outcome of area A cannot be explained only by the internal condition of area A. Using primary sources and analytical models, this paper demonstrates the effect of cross-area issue linkage in SII, and thereby overthrows conventional explanation on why U.S. pressure succeeded in some areas but not others.<br>The first part elaborates two spatial models of multi-issue bargaining between two governments, and suggests how the outcome of international negotiation is affected by whether or not participating governments can strike cross-area barters.<br>The second part empirically shows when the U.S. and Japanese governments were ready to make such barter deals. At first, the level of negotiation was sub-cabinet and both governments suffered from interagency conflicts, leaving no room for sacrificing one agency's issues for the sake of another's. This situation changed when top-level policy makers took direct command in both countries. The top policy makers in the U.S. allowed their Japanese counterparts to choose across areas, on the condition that the package as a whole is significant.<br>The third part quantitatively evaluates newly declassified records of the negotiation, and shows the timeline change in the Japanese responses to the U.S. requests. This clearly illustrates the reversal in the pattern of Japanese concession when the top policy makers intervened. At first, Japanese concession was concentrated in keiretsu and exclusionary business practice areas, where the potential targets of U.S. retaliation (export sector and large business) were most affected by U.S. demands. After the change in the level of decision making, significant concessions were made in the other areas, while the draft of keiretsu and business practices was left virtually untouched. This observed tendency fits in with the prediction of the models; however, no major preceding hypothesis on SII can explain it.<br>Theoretical research on bargaining has shown very clearly that issue linkages can completely change the bargaining outcome. However, empirical case studies of issue linkages in specific negotiations have seldom matched the clarity, owing to the difficulty in obtaining direct evidence of highly political barters, which involve intense conflict between interest groups, bureaucratic agencies, and politicians. By compensating the lack of direct evidence by theoretical models, this paper adds one clear example and demonstrates a method to do so.