A Common Value OTC Network Model

This thesis looks at decentralized client-dealer markets by representing them as a Bertrand competition over a network of connections and examines the Nash equilibrium bid-ask spread properties. We specifcally look at OTC financial markets with a common but unknown true value and show how equilibrium bid and ask prices can be viewed as the optimal amounts of signal reduction in a common-value first-price auction. We algebraically examine the equilibrium in the duopoly case and then analyze the equilibriums numerically in the n-dealer case. Our model suggests that changing transparency can have non-intuitive effects that depend on both the network structure and the information asymmetry. Similarly, the model can explain the empirically observed centrality premiums and discounts found in some OTC markets and other noted pricing anomalies. We apply our network model to the UK betting market using data from the UK General Election 2019. In betting markets, win probabilities are the asset, and dealers are the bookmakers, and we find that bookmakers' odds closely follow the model's predicted levels and exhibit a predicted centrality discount