This paper first estimates the number oftobacco users in 2000 and cigarette consumption from 1970 to2000 by regions and levels of development and brieflydiscusses the advantages and disadvantages of estimatingtobacco use on the basis of prevalence surveys or aggregatedata. Secondly, prevalence (and its associated number ofsmokers) and cigarette consumption (total and per capita)are projected in the future using several scenarios ofchanges in tobacco use (prevalence and cigaretteconsumption), as well as different assumptions aboutpopulation and income growth. The results show that even ifall countries immediately implement a comprehensive set oftobacco control policies, the reduction in the number oftobacco users and the total consumption of cigarettes willbe gradual. This should give comfort to farmers and otherswho fear the impact of tobacco control on their livelihoods.It is however, discouraging news for public health, since itimplies that the number of tobacco attributable deaths willcontinue to rise for decades to come.