than for lower-skilled workers, such as the desire for better matching of skills with job requirements among more important workers. Yet, firms in the sample are apparently not typically required to engage in an extended search for even a skilled worker. Of the 34 positions vacated by the deaths of skilled workers and later filled, 8 were filled within one week, and another 15 required only a second week of search. Based on these data, it is difficult to argue that the deaths of skilled workers will greatly impede the operations of sub-Saharan African firms.

It has often been conjectured that the loss of a small number of elite individuals in the economy can disproportionately disrupt economic and social activity. If the maturing cadre of younger leaders is too small or too inexperienced to fill adequately the roles of its deceased seniors, economic growth suffers. Table 6-3 presents data on worker attrition, but unfortunately these data do not permit a breakdown of attrition rates by skill category of worker. If, in the absence of AIDS, attrition among professionals and managers is much lower than the 6 to 12 percent attrition rates among the general work force, then a seroprevalence among managers of 45 percent would, according to Table 6-1, increase the mortality rate among this group by a factor of 10 (from 5 to 50 per 1,000). However, the results