ABSTRACT
One would think that, in the attempt to build and estimate empirical models of US and Soviet military expenditures, the choice of the data themselves would be relatively straightforward. In this chapter we demonstrate that this choice is not unambiguous. In so doing, we raise a number of data-related issues and questions which need to be addressed by modelers of a US-Soviet military expenditure arms race. At the outset, it should be noted that we focus our discussion on estimates of Soviet military expenditures, since these seem to be the ones which generate the greatest controversy. In addition, while our comments are aimed primarily at the research/modeling community, they could just as easily have been directed to the policy-making establishment. We believe that the plausibility requirements which we outline are equally applicable to both groups. By way of concluding we suggest that the empirical fit of one’s models (and, hence, the subsequent inferences which one is able to draw) is likely to be affected by the particular data series employed. Both modelers and policy makers recognize that an estimate of Soviet defense activities is of major importance. Why is the Soviet estimate so crucial? For one thing,
Moreover, as Abraham Becker, a senior economist at the Rand Corporation, has noted, considerable information can be conveyed easily in a summary fashion.
