Abstract: PPP contracts most often have durations of between 20 and 35 years, but in some cases even longer. The main reason for this is the wish of the Public side to minimize its financial contribution, by including in the contract many years of revenue generation by the project to help cover the investment contribution of the private partner. Implicit however is the need to fully amortize the initial investment, which in many countries is even included in the relevant legislation. PPP contracts are normally framed around the delivery of a range of services during the lifetime of the contract, those services requiring the initial construction or recovery of an expensive infrastructure. The specification of the financial clauses of the contract requires the estimation of demand for those services over the period of the contract and this is usually taken as the major incidence of uncertainty in the contract. Indeed, experience shows that demand forecasts often fail substantially, in many cases by more than 20%, mostly by excess, as State side project promoters (and the bidding private partners) tend to be excessively optimistic about the development of such demand. But when we consider the nature of these contracts we should recognize the existence of at least two other very important types of uncertainty: first, the socially desirable scope and specification of the services to be offered as technology and social preferences evolve; and second, the policy guidelines relative to the total quantity and the social distribution of those services, as that quantity may be causing congestion in other parts of the system, or it may become important to (positively or negatively) discriminate some user segments. In both cases, it is almost impossible to foresee at the time of writing the initial contract if, when and in what direction such types of socially beneficial changes in the provision of the services would intervene, but this rigidness may bear a great loss of social welfare in relation to a more adjustable framework. This criticism affects not only PPPs but all kinds of concession contracts with long duration, so it is not the “partnership” element that must be questioned but rather the duration of the contract. An alternative way is relatively straightforward: abandon the assumption that these contracts must provide full amortization of the infrastructure, which allows adoption of contracts with a shorter life, and the use of multiple such contracts over the lifecycle of the infrastructure. The first generation contract would still have to face the full cost of the construction, but the private partner would receive the unamortized part at the end of that contract, to be paid by the State, directly from the public budget if no more private participation is wanted, or indirectly through the acquisition fee for the contract to be paid by the partner to the second life segment. But, crucially, the State recovers the right to re-specify the terms of the service to be provided without the need for any indemnity, and also the uncertainty associated with the evolution of demand in that period will be much smaller, as this will be my then a mature system in operation. This may seem to increase the transaction costs for the State as more contracts (although of a similar type, especially from the second onwards) may have to be negotiated and signed. But if we take into consideration the difficulties of the frequently needed renegotiations of long duration contracts and the conditions of asymmetry of information in which the State normally finds itself in such cases, we will conclude that, besides avoiding the loss of welfare due to the poor fit of the contract after 20 years or so, this solution after all may also reduce the transaction costs associated with negotiations over the duration of the traditional contracts. [Copyright &y& Elsevier]