Beginning in the mid-1990s, the National Health and Welfare Ministry in India began a new phase of the overall campaign to bring better health care to India – the Reproductive and Child Health Programme (RCH) (Indian Ministry of Health and Family Welfare, 1998). One of the primary elements of this campaign is greater outreach to remote and poorly served regions. This outreach comes primarily in the form of establishing Primary Health Centers (PHC) and Community Health Centers (CHC) to serve relatively small population areas. These centers, in turn, can run RCH camps, either in their own location or in even more remote areas. In contrast to the PHC and CHC, the camps are mobile and temporary, usually lasting only one or two days, and offer only very basic health services (for a general discussion of these camps, see Mavalankar and Sinha, 1999).
There are two obvious concerns in identifying the effects in the paper based on variations in number of camps. First and foremost, there is the concern that the camps are targeted to areas based on existing vaccination conditions. A second, related, concern is that there could be selective migration to areas with many vaccination camps by individuals who have the most need for them (Rosenzweig and Wolpin, 1988). In this context, the latter issue is less likely to be a big deal because there is no reason to think that the same number of camps will be available in future years. So it is unlikely that people will choose migration decisions based on this year’s camp placement. The former issue is more of a concern. Both concerns, however, are addressed by the discussion in this section on the correlates of camp placement.
There appears to be no systematic scheme for camp placement. The RCH programme is run at the state level, suggesting that there may well be variations across states in the number of camps. In addition, since the camps were based out of PHCs and CHCs and travel is expensive, we may expect areas closer to these clinics to have more camps. Finally, village population and village area may also have a role, since the benefit of a camp is likely to be larger in bigger villages. Given these concerns, all regressions will allow for the effect of gender to differ between states and will include interactions between gender and village area, village population and distance to other sources of health care.
However, it is also important to consider whether empirical placement of the camps appears to be non-random beyond these controls. For example, are there more camps in richer or more well-educated areas, or areas with more children? To test for this, in Table 2, I regress the number of camps by village on some simple village characteristics (income, maternal education and age, village