Methods and Terminology
TO CONDUCT OUR ANALYSES, we combined publicly available data on Bay Area cities’ housing allocation numbers with publicly available U.S. Census Bureau data for median household income and racial composition. Using Pearson’s Correlation tests,7 we examined whether the demographic attributes were associated with the RHNAs for different income categories. Where historical data was available on permitting, we also used Pearson’s Correlations test to analyze whether there was any relationship between performance towards RHNA goals and city median income and city racial composition. Finally, we examined some of the cities which historically performed the best and worst relative to their RHNA targets.
In this report we utilize the following terminology (see the next column) to describe income levels and time periods.
Terms & Methodology
Area Median Income The California Department of Housing and Community Development annually determines affordability categories based on MSA-level8 median income data from the U.S. Department of Housing and Urban Development. Income categories are then constructed as a percent of the median income value, and RHNAs are developed based on the following percent ranges:9
Regional Housing Needs Allocation (RHNA) Cycles The time period in which goals for housing development are set. The housing allocations are assigned to cities once every eight years, a time period that’s referred to as a “cycle.
- 7. Pearson’s Correlation tests are used to determine the association between two variables, or in other words, the extent to which an increase (or decrease) in one variable is associated with an increase (or decrease) in another variable. Higher “r” values indicate that there is a stronger association; r values of (+/-) .1-.3 are considered weak, while r values of (+/-) .3-.5 indicate that the association is moderate and r > (+/-) .5 indicates that the relationship is strong. “P” values indicate how statistically significant a relationship is, or in other words, what the chances are that the association is random, rather than due to the observed data. High p-values indicate that there is a greater chance that the association is random, meaning that the findings are less credible.
- 8. Metropolitan statistical areas (MSAs) are geographies defined by the U.S. Office of Management and Budget and generally are comprised of multiple counties. For a map of MSAs in the Bay Area, see https://en.wikipedia.org/wiki/San_JoseSan_Francisco-Oakland,_CA_Combined....
- 9. While it is outside the scope of this brief to address this issue at length, it is important to note that cities which have significantly lower median incomes than their respective county’s median income receive fewer allocations for very low income and low income units than they would if RHNAs were tabulated using each city’s median income. The “area median income” therefore underestimates low income housing need in counties with large degrees of income inequality, since the median does not indicate the full range of affordability needs.