772
MARZIO GALEOTTI
        justified, thereby remedying the fact that it is the former which is relevant for the theory but only the latter is observed. The upshot of these developments is that also the q model is now a structural model with firm theoretical foundations.
             Both flexible accelerator and q models can be seen as stemming from a common underlying optimization framework where adjustment costs play a crucial role. In their most common specification, the models can be written as follows:
        (1)                             K = 6 (K* - K)
        (2)                             K = b(q- 1)
        where 9 is the (optimal) speed of adjustment of the capital stock to its desired level K*, and h is the (inverse of the) marginal cost of adjustment.
             By looking at (l)-(2), it appears that both models offer the same qualitative explanation of the investment mechanism-, this is put to work whenever a discrepancy between the target and the actual value of the key determinant is perceived. Comparing the two formulations, it also emerges that the disequilibrium that characterizes both models is looked at from a different perspective: the q approach underscores a price adjustment process \ while the flexible accelerator approach emphasizes a quantity adjustment mechanism. Furthermore, it also appears that the coefficients 9 and h perform the same function and deliver the same type of information: that of determining the pace at which investment optimally evolves as actual capital stock approaches its long-run desired level on the one hand, and as q reaches its equilibrium value (here taken for simplicity to be one) on the other.
             The analogy between the two theories has been noticed by Bosworth (1975):
             Although it appears much different, the securities-valuation model is conceptually identical to the version of the neoclassical model used in most empirical studies (p. 285).
        More explicitly, Tobin and Brainard (1977) observe:
             The hypothesis that investment is related to the difference between R (margi-
             This terminology is loose: this is a dynamic equilibrium situation, in the sense that investment evolves over time along an optimal path.
             This emerges more clearly from a formulation of the optimization problem in which adjustment costs are internal to the firm’s technology. In this case the q model has the following format: dK/dt = h (q — g), where q is the shadow price and g the market price of new capital goods. See Galeotti (1986).