Annual sinking fund modelling for housing association major repairs provision

This study carries forward the investigation of sinking funds in housing asset management. Under the new financial framework, established as a result of the 1988 Housing Act, Housing Associations are required to create sinking funds to finance future Major Repairs on all new and rehabilitated stock. The assessed maximum annual sinking fund contribution adequate for this is expressed in directives issued by Scottish Homes and the Housing Corporation. Mathematical programming models were developed, as an alternative to applying conventional financial calculations, to project sinking funds over a sixty year planning horizon for a number of new-build developments. Linear programming models were solved with both deterministic and stochastic maintenance data, in order to determine how robust a conventional deterministic life cycle costing model is for sinking fund projection. Maintenance programming is a dynamic problem and the timing of projected maintenance works will have to be reviewed periodically using information obtained from condition surveys. As a consequence the sinking fund strategy will have to be amended in the light of any new information. Using a dynamic model the extent to which these inevitable changes affect the original sinking fund strategy were investigated by simulating "actual" policy as it would evolve throughout the planning horizon. A number of conclusions are drawn using the results from the various models. Firstly, current assumptions on what constitutes an adequate level of annual sinking fund are likely to be inadequate to fund the long term Major Repair needs of new stock. Secondly, it is apparent that many diverse sinking fund strategies can be modelled that will fund a series of major maintenance expenditure at optimal cost. Thirdly, mixed-integer linear programming shows that under certain circumstances a conventional sinking fund, which is always in credit, can be inefficient. For some profiles of expenditure it may be more efficient, in terms of overall cost, if it is permissible for the fund to be overdrawn. Finally, the dynamic sinking fund model shows that making significant amendments to maintenance projections is likely to have an adverse effect on a sinking fund policy to the extent that many funds will be seriously underresourced. Therefore more accurate long term forecasts will reduce the likelihood of substantial changes having to be made to the sinking fund strategy in the future, minimising its cost.