This paper generalizes the (Mankiw, Reis, 2010) Neo-Keynesian model of partial information and aggregate supply. Firms observe the changes in the economic variables only imperfectly and need to form expectations about the market fundamentals. Unlike the conventional approach, which solves the model using first-order log-linearization, we allow for the inclusion of the second order moments of shocks. We find that higher dispersion in shocks results in a lower aggregate output. It does more so, the lower is the precision of the signal. Coupled with the (Mankiw, Reis, 2010) result that monetary policy is more effective the lower the precision of the signal, this creates a natural trade-off between correctional power of monetary policy and a permanent decrease in aggregate supply due to bigger uncertainty.