COVID's Impact on Data

COVID's impact on the economy can't be overstated. As such, the data from 2022 onward (which take COVID into account) are a radical departure from previous years.

The 2022 data is the first year of data with the new 2020 Census included fully as both a benchmark year and underlying the estimates. This means we have seen some big shifts when looking at the newest EASI data compared to earlier year releases, this is true not just with income but many Census-based variables since they have all been newly re-benchmarked.

You can see the household and family counts in San Francisco decreased quite a bit overall from 2020 to 2022, therefore this general decline is driving much of the family count declines you are seeing between 2020 and 2022 rather than income shifts per se.

If you convert the numbers into percentages rather than the actual numbers, this may be more helpful for you in looking at trends since overall changes in numbers make it harder to see trends when looking at counts. See attached Excel file and refer to the highlight columns at the right for an example of this.

Due to the timing of when data sources are released, the impact of the pandemic in data sources is not seen in 2020 but rather in the EASI 2021 data and in the EASI 2022 data in some cases, due to the lags in data sources. When there is an unusual shift such as the pandemic (other examples are natural disasters such as Hurricane Katrina), we are constantly updating our projections and modeling as the latest data comes out.

One impact observed by the unusual 2020 time was slowed income growth and this can explain slower growth observed in general in income in certain instances and in the higher income categories specifically. For example, one theory is it may have been “easier” for a family to shift into the $150-200K group during this time than to shift into the $200K+ group or especially the $500K+ group. In general with Covid times, we have seen some shifts in income down in 2022 with a then expected uptick back up for 2027 though perhaps more modest 5-year gains forecasted than was projected in 5-year forecasts created a few years back.

Another point to keep in mind is that the EASI income data for the highest income categories is indeed modeled data based off of the family counts and full income distributions integrating the various income data inputs and also trends over time. Therefore, there is some built-in uncertainty in each estimate and very small percentage shifts can also be impacted by normal variability in the distributions and estimates. This can be seen in particular in very detailed categories such as cross tabs by both children’s age and income.