||Private provision of public roads through build-operate-transfer (BOT) contracts is increasing around the world in both developed and developing countries. Under a BOT contract, a private firm builds and operates roads in a road network at its own expense, and in return receives the revenue from road tolls for a number of years. Then these roads are transferred to the government. Important issues are the length of the concession period, capacity, toll charges and the benefits to the private investor, road users and the whole society under various market conditions and government regulations. Several issues related to a BOT toll road project are studied. This thesis first develops a benchmark analysis of the BOT road contract viewed as a combination of concession period, road capacity and toll charge under the assumption that the demand and cost functions are known to both public and private sectors. A bi-objective optimization problem is proposed for maximizing social welfare and private profit. And the properties of the concession period, service quality, social welfare and profit gains along the Pareto-optimal frontier are studied. A variety of government regulatory regimes are also investigated. We prove that: the private sector tends to offer a lower road capacity and a lower service quality under the price-cap regulation, while it chooses a higher service quality, a higher capacity and a higher toll charge under the rate-of-return regulation than those associated with the corresponding Pareto-efficient solution. In contrast, we prove that both the demand and markup charge regulations lead to Pareto-optimal outcomes. User heterogeneity in value-of-time (VOT) has long been a fascinating issue in road pricing studies. This thesis investigates the effects of the user VOT heterogeneity on the properties of the Pareto-efficient BOT contracts. Under some technical conditions, we prove that, the Pareto-optimal road-life concession period is free from the effects of user heterogeneity, while, even with constant returns to scale, the service quality is not constant and dependent on the curvature of the mean residual VOT function. The effects of user VOT heterogeneity on the outcomes of various regulatory regimes are also investigated. We find that, both the demand and markup regulations fail to achieve the Pareto-optimal outcomes and result in a lower level of capacity and service quality. In the presence of traffic uncertainty, this thesis proposes the BOT contract with full and partial flexibility according to the instruments adopted by the public and private sectors. Full flexibility refers to the case in which the public sector promises an exogenous rate of return on the private investment and in turn can freely ex post adjust the contract in a socially optimal manner according to the observed demand curve. Partial flexibility refers to the case where the public and private sectors agree on an ex ante demand risk allocation by contract and the ex post contract adjustment can be made contingent on a Pareto-improvement to both parties. A preferred Pareto-improvement can be selected from the Pareto-optimal solution set of a bi-objective programming problem equipped with a rational preference. In comparison with the traditional rigid contract, in which, private sector burdens all project risk, the proposed flexibility of the BOT contract is valuable for the contract adjustment mechanism to solve the discrepancy between the public and private sectors. Two other issues are also investigated in this thesis. One is how to select the optimal combination of concession period, capacity and toll levels for a BOT toll road project if the yearly increasing operation costs are incorporated in the project. Another is how to choose the capacity and toll levels of an ”add-on” toll road parallel to an existing one with various ownership regimes, namely the existing road can be a free, public or private toll road.