This paper proposes a Markov chain model for studying the impact on asset prices of illiquidity associated with search and bargaining in an economy. The economy consists of finitely many agents who can trade only when they find each other, and any trade between agents changes the population of the agent types which affects the asset price in the future. Assuming that the equilibrium utility as well as the trade price is proportional to the asset dividend, we obtain the asset prices in steady state. Through extensive numerical experiments, we observe that the equilibrium prices exhibit the cutoff phenomenon (i.e. crash) as the fraction of pessimistic agents becomes large. Models with a market maker as well as irrational agents are also considered.