Short Selling Modeling Under Market Segmentation And Shorting Proceeds Restrictions

This paper has two parts. The first part introduces the security lending market and security shorting market. The second part, the main body of this paper, presents a new model with the assumption of shorting proceeds restriction, based on the working paper of Wang et al. In this model, I still assumed the existence of market segmentation, i.e. there are two groups of investors:optential lenders (investors type Ⅱ) who can only buy and short in the security trading market but can not lend their securities to short sellers, and liquidity traders (investors type Ⅰ) who can not only buy and short in the trading market, but also lend their securities to short sellers. Optimistic potential lenders are the only suppliers of short demand, and optimistic liquidity traders are the only factor that determines the security price in desired market. This model conbines the security lending market and security trading market, and determines the security price, lending fee and short interest simultaneously.This model is also made up of two parts. Ths first part is the perfectly competative lending market, which is presented based on the American market. In this part, this paper analyzes the influences on the equilibrium outcomes of each factor, including the degree of heterogeity, the population mass of potential lenders, investors’aggregate valuation, short selling restrictions, recall risk and budget constraint. The second part is the oligopolistic lending market under the framework of game theory, which is based on the Chinese market, the Japanese market and the Korean market. In this market, potential lenders will take part in a two-stage game. This paper analyzes the influences on equilibrium outcomes of the degree of heterogeity, investors’ propensity to risk, the population mass of liquidity traders, the individual valuations of potential lenders and budget constraint. Bacause of the complexity of game process, the influences of some factors are not so straightforward. Therefore, I desiged a specific numeric experiment for every factor, to illustrate the pattern and degree of the influence of every factor.