The financial crisis of 2007-2009 highlighted the importance of financial market frictions in the real-economy. There is a burgeoning literature that looks to the role of assets as collateral in over-the-counter transactions as a possible source of instability.  However, most models incorporate rational expectations and, as such, have difficulty in generating empirically plausible asset price features such as excess volatility, bubbles and crashes.  This talk will discuss the implications of introducing bounded rationality into search models that include a liquidity role for assets. Branch will begin by incorporating traders with heterogeneous beliefs who trade an asset in bilateral meetings. In this setting, there are four primary results. First, heterogeneous beliefs can, in general, lead to trading inefficiencies that would not arise under homogeneous beliefs. Second, heterogeneous beliefs can destabilize a stationary equilibrium and lead to periodic and complex dynamics that remain bounded around the steady state. Third, along these dynamic paths, there is a non-degenerate, time-varying wealth distribution. Fourth, dynamic equilibria can feature a time-varying extensive margin of trade. The unique theoretical predictions from heterogeneous beliefs provide a potential explanation of experimental findings reported in Duffy and Puzzello. Finally, a stock-pricing example will be presented that provides quantitative evidence on the importance of the liquidity channel for generating bubbles and crashes.