This article builds a model to forecast the number of newspapers that will exist in each US county in 2028, based on what is known about each county in 2023. The methodology is to use information known in 2018 to predict the number of newspapers in 2023. Having estimated the model parameters, we apply it to 2023 data. The model is based on market demographic characteristics and allows for different effects (slopes) for large, medium and small markets (population segments). While the main contribution is forecasting, we interpret the parameter estimates for validation. We find that the best predictor of the number of newspapers in five years is the current number of newspapers. Population size also has a positive association with newspapers. Average age and median income have positive slopes, but not in all population segments. The proportions of Blacks, and separately Hispanics, in a county have negative associations with the number of newspapers, but not in all population segments. The report provides maps showing which counties that are currently news deserts could be revived, which counties that currently have one newspaper are more at risk of losing it, and which counties with two or more newspapers are at risk. We also study the model residuals showing which counties are under- or over-performing relative to the market conditions.

14 pages