One of the major topics of the defense economics literature regarding Turkey and Greece has been the empirical modeling of various aspects of arms racing. However, despite a considerable amount of research, little evidence has been found in favor of an arms race between the two countries. In the literature, this failure of applied studies has been attributed, among other reasons, to the sensitivity of the results to the underlying model specification, to small sample size, and to measurement issues. This study uses novel, nonlinear, models to investigate the possible relationship between the military expenditures of the two countries. It is assumed that if there are two regimes characterizing the low (or negative) and high-growth military expenditure periods, the growth rates of one country's military expenditure may have distinct effects on the military expenditure regimes of the other country or may contribute to the change from one regime to another. The nonlinear models examined are Smooth Transition Regression models (STRs). Strong evidence of nonlinearity for Greece is found, with asymmetry relating to two distinct regimes through lagged Turkish military expenditure changes.