Stability of Radner equilibria with respect to small frictions (Q1709608)

The authors consider a general equilibrium model with transaction costs. They assume a model with heterogenous agents. Two agents participate in the financial market and the third agent act as an operator of an exchange receiving trading fees. The endowment of agents and dividends from the financial assets are modelled with diffusion processes (stochastic differential equations). The agents divide their endowments into consumption and investments into a risky asset choosing their strategies (investments-savings plans). Each agent maximizes his expected present value of utility from a stream of consumption in a finite time horizon. The authors look for a Radner equilibrium in this model, i.e. for such a combination of price processes, investment processes and consumption processes that both stock market and good market clear and the investment-consumption choices of agents are optimal. It is known, that if there are no transaction costs, such an equilibrium exists and the authors recall this results. To consider the existence of an equilibrium in the situation with transaction costs, the authors define a concept of asymptotically optimal strategies. They are investments-savings plans that minimize the performance loss compared to the frictionless case at the required leading order \(\varepsilon\). The asymptotic equilibrium with small transaction costs is defined as a situation when market clears and agents' strategies are asymptotically optimal. The article contains two main results. The first one is Theorem 3.2. It states that if interest rate is constant and some technical assumptions are fulfilled, then there exists an asymptotically optimal strategy at the leading order \(O(\epsilon^\frac{2}{3})\). The authors present a form of such strategy. They prove the theorem by approximating optimal strategy by series of auxiliary strategies defined as stopped processes. The second important result is Theorem 4.1, which the authors call ``stability theorem''. It states that the equilibrium in the frictionless market can be approximated by asymptotic equilibria with small transaction costs. At the end the authors give an example of a benchmark model of economy, in which endowments of agents given by stochastic volatility models. The authors take very specific assumptions on the volatilities of volatilities (there should sum up to zero across agents). The authors present investments and savings plans in equilibrium in this benchmark economy.