This paper explores the relationship between lease length and the allocation of responsibilities for leased premises. Longer leases would be expected to give tenants more control over their premises than short leases. This paper explains the reasons for this, describes observed variations in lease structures in several markets and investigates whether there might be threshold lease lengths, above which major responsibilities are or should be shifted to the tenants. Typical lease lengths for non-residential properties are changing in some European countries, either as a result of legislation or market forces. Therefore, it is timely to consider whether the liabilities under these lease are likely to be or should be allocated differently in future. It is generally accepted that tenants do not have the same incentive as owner-occupiers to care for their buildings. Many writers have recognised that leased premises are likely to be neglected because lease covenants are only enforceable at significant cost. Less attention has been given to the potential neglect by landlords who have granted long lea161es or to the potential overspending by landlords recovering costs by way of service charges. Further, most of the literature is not explicit as to how the length of leases influences both the likelihood of neglect and the lease covenants that might encourage proper care of the premises. This paper summarises and uses a mathematical model of the costs of leasing in order to compare different ways of allocating duties. The model shows, not surprisingly, that the length of the lease is a major determinant of which party should look after the property. However, there appear to be two other factors ñ economies of scale and willingness to carry risks - that also affect the allocation of responsibilities and are largely independent of the length of the lease. International comparisons confirm that, where longer leases are used, tenants are usually given more responsibilities for the premises. Because convincing empirical tests of the relationship between lease length and the allocation of responsibilities are largely impractical, the effects of changing length on the costs of leasing on several common bases are simulated. The simulations, based on the mathematical model, estimate the lease length at which landlords and tenants may benefit from switching responsibilities for the property under a range of different scenarios.