By Adam Pagnucco.
Since 2006, I have written more than 4,000 posts about state and county politics and government. Few of them have had more consequences than County Government: Census Building Permit Data is Wrong.
That post revealed that census building permit data showing a crash in county homebuilding was in fact not based on data reported by the county, which had not been sending it to the Census Bureau for years. Since then, a county investigation has yielded more findings.
The Census Bureau is supposed to collect data from local building permit offices, to which it then applies imputed estimates and reports on a monthly basis. There are three issuers of building permits here: the cities of Gaithersburg and Rockville, which issue permits inside their municipal boundaries, and the county government, which issues permits elsewhere inside the county. Census told our county government the following.
*****
For the Building Permits Survey (BPS), we collect data at the jurisdiction level, and our county-level estimates for Montgomery County are made up of 3 jurisdictions – Rockville, Gaithersburg and Montgomery CO. unincorporated area.
Rockville looks like the best and most consistent reporter to the BPS. We receive their report nearly every month. They reported multi-family units pretty consistently up through 2019 and then between 2020 and 2022 have reported no multi-family which may be something they want to verify is correct. Our contact for Rockville is [redacted] and we attempt collection via email to this contact every month.
Gaithersburg stopped reporting to the BPS in April of 2020. When they were reporting, they consistently reported multi-family units prior to 2014 and reported no multi-family units between 2014 and April of 2020 when they stopped reporting altogether. Our last contact for Gaithersburg was [redacted], and we continue to attempt collection via a mailed letter every month.
Montgomery County unincorporated area has not reported directly to us consistently since 2012. Our last contact for Montgomery County unincorporated area was [redacted] and we have attempted collection via email to this contact every month. Just this morning, we received a phone call from Rabbiah Sabbakhan, Director, Montgomery County Department of Permitting Services who intends to provide a new point of contact for BPS to help us resume getting reported data.
When a BPS report is not received, missing housing unit data are either (1) obtained from another source (e.g. Survey of Construction (SOC), third-party data), or (2) imputed. If data are not reported and are not available from other sources, estimates are imputed based on the assumption that the ratio of authorizations for the current time period to the prior year total is the same for reporting and non-reporting jurisdictions in that Census Region. Data for the four types of structures (one unit, two units, 3-4 units, and 5 units or more) are imputed separately. Imputation of single units is further separated by metro status.
Obtaining good contacts for these jurisdictions and developing a consistent reporting relationship is likely the best way to help resolve any concerns on the quality of these BPS estimates. We are hopeful that the new MoCo unincorporated contact will help, and we will continue to pursue alternative contacts for Gaithersburg. In addition we hope to work closely with the contacts for these jurisdictions to ensure that multi-unit residential projects are being accurately reported to the BPS as multi-unit residential as opposed to commercial. If you have any additional questions, please don’t hesitate to reach out.
*****
The Elrich administration has done good work to restore our relationship with Census. Assistant Chief Administrative Officer Jake Weissmann and county spokesman Scott Peterson deserve credit for identifying this problem and taking action to fix it. That said, this issue has many tentacles.
First, this goes far beyond our county. Since my original post, I have heard stories of misreporting in other jurisdictions. One planner in Upstate New York told me that his county had 45 different permit-issuing jurisdictions. His state government asks local jurisdictions to report the number of permits they issue, not the number of residential units permitted. Since multi-family buildings contain more than one unit, those numbers are very different. Accordingly, he believes that at least some jurisdictions report the number of permits to Census just as they do to the state, creating a huge measurement problem. It’s not clear if Census has a way to detect such a problem or correct it.
An even bigger issue is that the building permit series is an input to other statistical series. Two of them are the American Community Survey, which is Census’s primary local demographic series between decennial census years, and its annual estimates of population. In April, I reported that Census estimated our county had lost population two years in a row. Was that calculation based on erroneous building permit information? What if the inclusion of county permit data were to cause those estimates to rise instead? It’s hugely important for county policy makers to understand whether we are gaining or losing population.
Building permits are also an input to national income and product accounts, including gross domestic product, measured by the U.S. Bureau of Economic Analysis (BEA). Now those series contain MANY inputs besides building permits so the impact of flaws in that data may not be severe. But it’s still unhelpful.
There are other consequences to bad data. Some federal and state aid programs are based on population. Are we getting less federal and state money than we should because our population estimates are pushed downwards by Census underestimates of building permits? If so, how much is that costing us?
Finally, there is the issue of private sector usage of building permit data. We have no idea how many universities, investment firms, economic consultants and other influential players across the nation are using this data to form opinions about our economy. The state is definitely using this data because the Maryland Department of Planning publishes it on their website. How many of these players think our homebuilding industry is an economic black hole when in fact it may be stronger than Census estimates show?
There is still a host of data showing that our local economy has challenges. For example, building permits are not an input into employment and establishment series maintained by the U.S. Bureau of Labor Statistics. But as someone who has been measuring our local economy for more than 15 years, I feel like I know less about our county than I did a month ago.
The Elrich administration did not create this problem and they deserve praise for trying to fix it. But they now have a huge challenge. They must audit their own data to ensure that it’s accurate. (I have heard questions about that issue.) They must recreate a reporting relationship with Census. They must help Census fix the historical building permit data going back many years. And they must push Census and BEA to fix any other series in which building permits are used as an input. It’s a massive job, but if we are going to accurately measure our county, it must be done.