A central element of the National Health Insurance Law that came into effect in 1995 was the removal of all barriers to the free movement of citizens between sick funds. One of the aims of the removal of barriers was to eliminate the possibility of "cream skimming" - the deliberate recruitment by sick funds of only relatively healthy new members. This article examines whether cream skimming existed in the Maccabi Health Fund before 1995. In the absence of health status data it uses data on health expenditure as an indicator of health status and compares the cost of new members before 1995 to the cost of new members since 1995. In addition, a comparison is made between the cost of new members and the cost of immigrant new members (no restrictions were put on immigrants joining Maccabi before 1995, and they can be used as a crude control group, free of any cream skimming). The existence of cream skimming should be reflected in higher costs of members joining after 1995, compared with those joining before 1995, and in higher costs of immigrant new members compared with those of non-immigrant new members. The results are mixed. For certain age groups, the data are consistent with cream skimming. However such data are also consistent with adverse selection - the tendency after 1995 for citizens of poorer health to join Maccabi - and it is difficult to separate the impact of cream skimming from that of adverse selection. What is clear is that there was not a jump in the number of new members above age 60 joining since 1995, indicating that the policy of cream skimming in Maccabi appears to have principally taken the form of restricting the recruitment of the over 60's - healthy and not healthy alike.