Improving data collection on housing production
* What challenges do local civil servants face in gathering and submitting the data? How long does the process take each year? What resources do jurisdictions have access to and use, e.g. software, consultants, etc.?
*How could the state improve APR?
Background
The California Department of Housing and Community Development (HCD) collects annual housing production data to track housing element implementation and progress toward achieving Regional Housing Needs Allocation (RHNA) goals. The Housing Element Annual Progress Report (APR) data, which is self-reported by cities and counties, is presented in a dashboard and can be easily downloaded on HCD’s website.
Researchers, as well as housing advocates, use the APR data to study housing production in California. The data has the potential to be a valuable resource, given its level of detail and potential for comparison across jurisdictions. However, the data submitted by localities is often inaccurate or imprecise. Planners or other local government employees are responsible for record-keeping and reporting on all proposed housing developments, but data management is likely not a top priority or area of expertise for most local government staff. While there is software that can help streamline data management and reporting, many cities, especially those with limited resources, do not have access.
A former civil servant who managed the APR data submission for a mid-sized county highlighted several challenges that made the process complex, onerous, and time-consuming. Specifically, there is significant potential for inaccurate duplication of entries, and civil servants must be diligent in monitoring for and removing accidental duplicates – while also being sure to keep entries duplicated appropriately per HCD instructions. Local government staff must also be careful to assign entries to the correct table in the submission file.
As a result, when the APR data from all jurisdictions are combined, the result is a noisy dataset of questionable accuracy. Researchers at the Terner Center for Housing Innovation, who have analyzed APR data for reports on topics such as the application of Senate Bill 35, advised our team that the data must go through considerable cleaning before any analysis can be performed.
Researcher Moira O’Neill has also identified issues with the APR data. When comparing the APR data San Francisco submitted to HCD for the years 2018 to 2021 with their own rigorous dataset, O’Neill and coauthors find that about half of the entitlements in San Francisco’s APR data (94 entitlements) were reported incorrectly, for example, with the incorrect year. Additionally, San Francisco failed to report 54 entitlements issued during that period that were identified by the researchers.
Description of Proposed Project
As part of the Abundance Accelerator initiative, the Possibility Lab is partnering with HCD to improve the quality and usability of the APR data. As we advocate for increased housing production, it is crucial that interested parties have access to reliable data on entitlement and permitting.
To understand the specific limitations of the APR data and how it might be improved, our research team is holding conversations with researchers and housing advocates to learn what data would be most useful to them. Additionally, we will interview local civil servants from different cities and counties to understand how data reporting processing could be more efficient and less error prone.
This project seeks to collaborate with HCD staff to design and implement a solution that addresses the main challenges identified by researchers and civil servants, as well as any concerns of HCD staff. This solutions may include options to:
- Improve the APR submission process by developing new systems, processes, forms, or reference materials to help local governments track and report data more efficiently.
- Streamline the reporting process to eliminate unnecessary duplication and minimize opportunities for human error.
- Develop additional processes for data cleaning and validation checks that can be performed by HCD staff or that are performed automatically at the time the data is uploaded. This might include automatic checks to ensure that responses are logically consistent.
- Improve usability and accessibility of APR data available on HCD’s open data webpage, such as through additional training or tools to help users filter for and download subsets of the APR data.
Expected Contributions to Research and Practice
Making the APR data reporting process less cumbersome would benefit local government employees by saving them valuable time and effort. This is especially true in small, under-resourced jurisdictions with limited staff capacity and out-of-date technology. Streamlining reporting could also improve the quality of the APR data by reducing opportunities to introduce human error.
Researchers can use the improved APR data to precisely understand the kinds of housing projects being approved by cities and counties and the processes they must go through to be approved. What kinds of entitlements must projects obtain, and how long, on average, does it take for projects of various sizes or in different locations to move from application to certificate of occupancy?
It also might be interesting to know if reporting errors, taken together, are random or if they vary in the same direction (or to the same degree) across jurisdictions. For example, this might allow us to find that jurisdictions, on average, tend to overstate or understate their production of low-income units.
The answers to these questions could inform public policy at the local and state levels. With its enforcement authority, HCD could direct cities and counties to implement policy changes uniquely targeted to address the jurisdiction’s specific deficiencies and roadblocks to building housing.
