Forecasting a Tribal Population Using the Cohort-Component Method: A Case Study of the Hopi

Abstract

Population forecasting is a difficult endeavor. When it involves a “small” population, forecasting becomes even more difficult because of the lack of adequate input data to appropriately implement a technique preferred by demographers, the cohort-component method (CCM). Small populations also are subject to high levels of stochastic uncertainty, which can lead to substantial temporal fluctuations in size as well as mortality, fertility, and migration rates over short periods of time. When the population is further defined in terms of characteristics such as tribal membership requirements, the acquisition of relevant age-sex and component data for the discrete population adds even more difficulty because the usual data sources may not have precise data for a given tribe, especially when there is a need for long-term forecasts as is the case in disputes over water allotments. In this case study of the Hopi tribal population, I show that these problems can be overcome, resulting in reasonably accurate forecasts over a long period of time. To do this, I identify three key input data sources: (1) an historic tribal census that serves as the basis for tribal membership; (2) the Social Security Administration’s life tables, which are specific to cohorts by year of birth, staring in 1900; and (3) an accurate annual record of tribal membership record that covers a time period that can be used as a benchmark once the forecast is launched from the historic tribal census. For this case study. these three data sources allow me to launch an 80 year CCM forecast of the Hopi Tribal Population from its 1937 tribal census to 2017 and conduct an ex post facto assessment of the accuracy of the forecasted total population in five-year increments from 1992 to 2017 using the annual tribal membership roll for this period; the assessment finds that the forecast matches up well with the tribal membership roll. So without adjustments, the forecast is taken out 20 years more, from 2017 to 2037. The results suggest that accurate long-term forecasts are possible using the cohort-component method for the Hopi, which, in turn, suggests that this same CCM process could be used to develop reasonably accurate long-term population forecasts of other tribes for which a tribal census was conducted that is linked to current tribal membership rolls, especially when these forecasts are required in resolving disputes over water allotments.