The dissertation develops a theoretical model to examine the effects of limited liability contracting on learning-by-doing and capital investment within a new agricultural industry. The theoretical model applies to many new bio-based industries, where novel crops are being used to produce goods, such as chemicals and energy, which would not be considered traditional agriculture.

Limited-liability contracts create an environment of moral hazard in learning investment and adverse selection in the production of the intermediate good. These two features of the contracting environment present difficulties for the principal to benefit from the learning-induced cost reductions realized at the intermediate stage of production. Thus, the principal under-invests in the industry and requires less of the intermediate good. Reduced feedstock orders decrease the incentives for the agent to invest in learning, and so the ultimate cost of production of the intermediate good is higher than optimal.

The dissertation adapts the theoretical model to construct a simulation of investment and production decisions within an industry for the generation of electricity using biomass. The results of the simulation show that an industry formed around limited liability contracts realizes project scales 25-30% smaller than optimal. Learning-induced cost reductions in the production of biomass are 20% less than predicted by engineering analyses. Limited-liability contracts raise the price paid by the principal for the feedstock by 25% above optimal. The analysis reveals that the price of electricity necessary for a project to break even is 5% higher under limited liability contracts. Sensitivity analysis illustrates that the problem of underinvestment increases under conditions favorable to grower learning.

A capital subsidy paid to processors that invest in technology encourages over-investment in capital relative to feedstock utilization. The Renewable Energy Production Credit or a feedstock subsidy paid to growers increase project scales by about 30%, yet they are still 20% smaller than optimal. These subsidies do not have a significant impact on the price of the feedstock to the processor. The government may seek to explore policies that encourage forward vertical integration in the industry.