The problem of national accounting “at constant prices” is in fact a problem of comparability of time series, as changes in the price structure preclude any direct comparison of economic flows. If such accounts are established they will make it possible directly to compare the same flow at two different times in the economy as a whole, and this without leaving the influence of other flows out of account. This makes it possible both to synthesize and to undertake analytical comparisons. The accounts could then be used for the study of time series, for projections or for structural studies (e.g. the mechanisms underlying the changing pattern of income distrubution).The first part of this report sets out to study the main problems of compiling accounts at constant prices and to examine what conventions should be adopted.The second part of the report considers how productivity gains can be explicitly shown in the national accounts. The proposed study plan restores the symmetry between price and productivity. As in the accounts at constant prices, gap variables are introduced to measure productivity gains. These variables can be interpreted in terms of surplus; the concept of surplus used here, however, is not the one adopted for the accounts in constant prices, but its dual. Setting up an accounting system “at constant productivity” therefore makes it possible to complete the information provided by an accounting system “at constant prices.”These two systems can of course be integrated: this leads to the introduction of the concept of an accounting system “at constant prices and constant productivity.” Such an accounting system makes it possible to show, in the same accounting framework, the respective contributions of price changes and improved productivity to the gains realised by the different economic agents. It therefore gives a complete picture of “transfers” between the agents. At the same time, the data on price and productivity can be integra