The possibility of indeterminacy and sunspot fluctuations in dynamic rational expectations models has been often questioned on empirical grounds, for such models are widely believed to rely on implausibly high degrees of increasing returns to scale and/or other controversial calibrations of economic fundamentals. In this paper, we study the occurrence of such phenomena in a standard (one-sector) optimal growth model with endogenous labor supply and a partial cash-in-advance constraint on consumption purchases. We show that, under standard preferences and constant returns to scale in production, indeterminacy typically prevails for an arbitrarily small amplitude of the liquidity constraint. We also analyze the cyclical properties of the model submitted to technological and beliefs disturbances and observe that it performs as well as comparable indeterminate models in the literature. JEL Classification: D90, E32, E41.

Standard stochastic growth models provide theoretical restrictions on output decomposition which can be used to investigate whether productivity shocks played a major role in observed business cycles. Applying these restrictions to US data leads to the following findings: (i) Business cycles implied by productivity shocks are mildly correlated to overall fluctuations and help account for a few episodes of US postwar recessions. However, only 20% of US fluctuations can be explained by these shocks. (ii) Most fluctuations seem instead to be due to “nominal demand” shocks, i.e. shocks which move output and prices in the same direction, but whose effects on output are ultimately transitory. (iii) Canonical sticky price models in the new-neoclassical synthesis tradition can account for the cyclical comovements of output and prices, but canonical, frictionless, RBC models cannot.