In Great Britain the demands for increased public construction investment have been increasing. However, although it is necessary to establish the amount currently being spent before being able to consider the need for an increase, this is difficult as there are two disparate sets of data available on public construction investment. Expenditure data are collected from public sector clients who commission construction work and output data are supplied by the contractors who undertake the work. Between 1976/7 and 1984/5 the expenditure data exceeded the output data by an average of 81.4% and the percentage difference has been increasing. This paper identifies the causes of the differences between the two sets of data and where possible quantifies them. Even after amending for these quantifiable differences the expenditure data exceeded the output data by an average of 10.5% for new public housing work and 29.4% for new public non-housing work. As the remaining unquantifiable differences are clearly significant this paper concludes by examining possible future methods of compiling accurate data on the level of public construction investment.