Household saving, which is a part of national saving, is expected to contribute to economic growth significantly. Its share in the total national savings of Pakistan, during 1960-90, was about 83 percent, ( and in gross national product it varied between 6.6 percent and 10 percent, in the early 1980s.2 This is a significantly high proportion considering the meagre total national saving ratio in Pakistan. The analysis of savings is a controversial issue. First. there is no standard empirical definition of household savings. Should consumer durables be considered a part of household saving? Should human capital be considered a part of household saving? These are important questions. Answers to these questions will be useful for the analysis of consumer preferences and for public policy. The rationale for the inclusion of consumer durables as a’ part of household saving is that they are like productive assets purchased in order to provide for a flow of services. Similarly, human capital is expected to raise labour productivity and its future income and consumption. The second issue is the response of each component of saving to changes in economic and demographic factors.3 For example, jewellery and assets may respond differently to changes in economic and/or social conditions. Third, since the theoretical literature defines saving as a residual there is no standard functional form of the empirical saving function. In this paper, we estimate different functional forms to analyse savings behaviour in Pakistan.