Stochastic Dynamics and Matching in the Old Keynesian Economics: A Rationale for the Shimer's Puzzle