MODELING THE US-SOVIET ARMS RACE AS AN ERROR-CORRECTION PROCESS Sam-man Chung National Defense Staff College, Seoul

If two states are engaged in an arms race, then we expect to find their armament levels moving together over time in equilibrium. Such a “moving equilibrium” in US and Soviet armament levels -or cointegration, in econometric parlance -represents a commonsense understanding of the superpower arms race. Most empirical investigations have started with this commonsense understanding and have sought to reveal the short-run action-reaction dynamic that generates this moving equilibrium in military expenditures. But to hypothesize such an action-reaction process is to expect rather much from lessthan omniscient policymakers and their lumbering state bureaucracies, and the empirical literature offers little evidence to the contrary. The preoccupation with short-run dynamics also neglects the important long-run dimension of the arms race. If the minimum condition for the existence of an arms race is the co-movement of two states’ military expenditures over time, then it must be the case that this long-run equilibrium is maintained by one or both. Time-series analysts have termed such behavior error correction. Our analysis indicates that six pairs of time series -Soviet military outlays and US service requests, presidential requests, and congressional authorizations in both current and constant dollars -are cointegrated, suggesting that US and Soviet armament efforts maintained a long-run equilibrium relationship. That is the pattern we usually associate with an arms race. But the relationship was not symmetrical, for it was the US policymakers and not the Soviets who adjusted their defense budgets in the face of deviations from equilibrium. The superpower arms race appears from this evidence to have been one sided: the Soviets acted, the Americans reacted. Relative to these long-run dynamics, evidence for US reactivity in the short-run is sparse. Overall, the impact of Soviet outlays on US defense budgeting was mostly indirect -i.e., to the extent that Soviet policy caused deviations in the equilibrium relationship, which US policymakers sought to correct.