Threshold Behavior and Aggregate strongFluctuationstrong - Makoto Nirei.pdfVIP

Threshold Behavior and Aggregate Fluctuation Makoto Nirei Utah State University, Department of Economics, 3530 Old Main Hill, Logan UT 84322, U.S.A. Abstract This paper concerns a propagation mechanism in an economy where many indi- viduals follow a threshold rule and interact with a positive feedback. We derive an asymptotic distribution of the propagation size when the number of the agents tends to infinity. The propagation distribution exhibits a slower convergence to a deterministic value than it would if the agents followed a smooth adjustment pol- icy. This gives rise to significant aggregate fluctuations in a finite lumpy-adjusting economy even when the agents are hit by small independent shocks. Key words: Aggregation, propagation, cascade, law of large numbers, heavy-tailed distribution, (S, s) policy, lumpy investment, menu cost, strategic complementarity, best response dynamics JEL classification: E1, E3 1 Introduction Since the debate between Prescott and Summers [1,2] on aggregate techno- logical shocks, it has been of continuous interest whether idiosyncratic shocks can cause aggregate fluctuations. Whereas idiosyncratic sectoral shocks have robust tendency to cancel out each other in the standard dynamic general Email address: mnirei@econ.usu.edu (Makoto Nirei). 1 This paper is based on my Ph.D. dissertation. I am grateful to Lars Hansen, Fernando Alvarez and Jos´e Scheinkman for their advice. I have benefited from com- ments by Samuel Bowles, Michael Chwe, Doyne Farmer, Xavier Gabaix, Jess Gas- par, Luigi Guiso, Dana Heller, John Leahy, Toshihiko Mukoyama, Andrea Tiseno, and anonymous referees. This work is completed during my residency at the Santa Fe Institute. Preprint submitted to Journal of Economic Theory