2022 Simulation

In the fall of 2025 OFM met with county and regional planners to show the results of a projection simulation.  The simulation took the data inputs used for the official 2022 projections, as much as practical, and applied them to a new county-level probabilistic projection model to demonstrate how the results of that model compare to the official 2022 projections.  OFM is considering using the probabilistic projection model in 2027 instead of the more deterministic model that has been in use since 2002.

Presentation slides for all counties: Probabilistic Population Projection Model: In-Development

Feedback:

The feedback from the meeting was very positive.

OFM asked for feedback on the following questions.

• Q: Do explicit probabilistic claims make a difference for your work?
  A: Explicit probabilistic claims were viewed as positive but not required.

• Q: Would the estimates exceeding the bounds in the five years before the next projection be problematic?
  A: No. There is no penalty or issue with a county’s population estimate exceeding the bounds in the short term. GMA plans are intentionally long-term plans and would not be impacted by some short-term variability.

• Q: What are your thoughts on the width of the bands (we presented a couple options for prediction intervals)?
  A: Participants expressed cases for bands that are narrower and wider at the projection horizon than the bands in the simulation. As expressed, OFM is currently leaning towards 90% interval bands.

• Q: What are your use cases for the age-sex data?
  A: Some participants use the age-sex data for planning and simulations including transportation planning and housing needs planning.

• Q: Would adding another five to ten years to the official forecast horizon be helpful?
  A: Yes, a longer projection horizon is helpful especially in the latter half of the decade. A longer horizon is also helpful/necessary for transportation planning.

Participant questions:

• Q: Do OFM’s projection account for climate change, climate related migrants/refugees?
  A: OFM’s model does not have a specific climate adjustment factor, however, climate change has already been happening, and people are including climate in their migration decisions. To the extent that climate has been a factor in historical migration choices, it is captured in our historical data and thus impacts possible migration futures.

Resources

• bayesPop website: https://bayespop.csss.washington.edu/
• Yu, C., Ševčíková, H., Raftery, A.E., and Curran, S.R. (2023). Probabilistic County-Level Population Projections. Demography, Vol. 60(3): 915-937. (https://doi.org/10.1215/00703370-10772782)
• Ševčíková, H., Raymer, J., and Raftery, A.E. (2024 preprint) Forecasting Net Migration By Age: The Flow-Difference Approach stat ArXiv, Vol. 60(3): 915-937. (https://doi.org/10.48550/arXiv.2411.09878)