Synthetic annual place-level estimates of individual and household employment status, earnings, income, and poverty status for American Indian and Alaska Native individuals in rural Alaska, 2000 - 2016

The American Community Survey (ACS) replaced the United States (US) Census Long Form Survey after 2000 as the only systematic source of community-level information on the American Indian and Alaska Native (AIAN) population living in rural Alaska communities. We define rural Alaska as the region corresponding to Census Public Use Microdata Area (PUMA) 400, known also as the Subsistence Alaska PUMA. The ACS is an annual survey. However, due to small sample sizes, the ACS publishes estimates for places and census areas/boroughs only as five-year moving averages. Even with the averaging over five years, estimates of economic conditions vary substantially over time in many communities, and high margins of error make it difficult to distinguish communities from each other. Using individual Census and ACS records accessed through the Census Research Data Center program, we estimated censored and logistic regression equations with random community effects that explained individual survey responses for year-round employment, individual and household earnings, individual and household income, and poverty status for individuals indicating AIAN identity, either alone or in combination with other races. Explanatory variables in the equations included Alaska Department of Labor place-level employment and earnings, Alaska Permanent Fund Dividend payments, and characteristics of individuals and households such as age, gender, presence and number of children in the household, etc., that can be benchmarked with precision to the 2000 and 2010 census counts. Annual predicted values of the equations averaged over individuals in each community were combined with survey means adjusted for random interannual survey age and gender variation to produce empirical Bayes synthetic annual estimates with margins of error comparable or smaller than the published five-year averages. The synthetic estimates include the regression-based predicted values for years 2001 through 2004 when no survey data were collected. Data are restricted by the US Census Bureau to avoid disclosure of individual information until approved for public release. The data are still under disclosure review, and will be provided if released.