Abstract: Abstract: Technological innovation is a complicated system which is characterized by the uncertain products and market as well as irreversible investment. As a method of investment valuation under the uncertainty, a real option is widely applied to the investment of technological innovation. The fact that many firms deploy the principal-agent structure when making investment in technological innovation inevitably raises the problems of information asymmetry, such as the moral hazard and adverse selection of agents. The adverse selection model of the principal-agent framework is applied to the real option valuation of technological innovation investment. First, a real option valuation model is setup under symmetric information, which gives the boundary conditions of investment. In the case of information asymmetry, it is assumed that all of investment cost and part of investment benefit are in public to the both principal and agent while the rest of the benefit is only known by the agent. Under this premise, an adverse selection model with a continuum of possible types is developed by building the profit function based on the real option valuation. Consequently, the model gives the optimal contract of the agent and the investment boundary conditions of the principal. In the end, an example is taken to demonstrate the economic implication of the real option model under asymmetric information. Also, the two models developed under symmetric and asymmetric information are compared, from which, two findings are obtained. One is that all types of agents except the one with the highest productivity have a lower boundary value for the investment under asymmetric information than they do under symmetric information. The other is that all types of agents except the one with the lowest productivity are able to harvest a strictly positive information rent.

Key words: investment in technological innovation, real option, asymmetric information, adverse selection, continuum of types