Philly Demolitions
In January 2021, a bill called the “Mixed Income Neighborhoods Overlay Bill” was passed into law in Philadelphia requiring a minimum percentage of affordable housing units to be set aside for new developments of 20 or more units. A form of “mandatory inclusionary zoning”, the goal of the law is to increase the supply of affordable housing in the city and prevent displacement of local residents. Notably, this law only applies to certain sections of the Third and Seventh Council Districts, areas that council deemed residents to be the most vulnerable to displacement. This is no surprise, considering that the law was spearheaded by Councilpersons Jamie Gauthier and Maria Quinones Sanchez, of the third and seventh districts. This law builds on legislation passed on previous years that offered incentives for inclusionary zoning, but no mandates.
With home prices increasing steadily over the past few years- nine percent alone last year, according to Zillow - many argue that the only way to ensure that locals are not displaced is by implementing policies like the Mixed Income Overlay bill. It would be unrealistic to expect development to come to a full halt, but proponents of these mandatory inclusion zoning policies argue that they allow for continued development and growth of neighborhoods and tracts, while holding developers accountable for providing at or below market rate housing.
However, a law like this works best when it is targeted at the tracts that are experiencing the most new development- something that is often talked about interchangeably with gentrification. Development in of itself is not always a negative, as economically disadvantaged neighborhoods could benefit from development that improves the quality of housing stock, reduces vacancy and property decay, or brings in in new businesses. Similarly, Councilmember Gauthier was quoted by WHYY saying: “..the continued growth of our city is important, but it is unacceptable that it comes at the detriment of vulnerable Philadelphians…with many developers prioritizing profits, regardless of the social repercussions, the best way for us to ensure that affordable housing options remain available…is policy change”.
With this in mind, does the Mixed Income Overlay Bill make the correct call in the areas it encompasses ? Are these the best zones to protect at risk residents, and also areas that are experiencing enough recent development to give the law some teeth? Requiring affordable housing unit minimums doesn’t have nearly the same impact if there is not actually significant new development in the specified area- in this case, of 20+ unit buildings, which is the minimum unit count for a development to be included in mandatory minimums.
It is unclear what methodology was used to select these specific zones in the third and seventh districts. According to Nigel Thompson of Al Dia News Media, last year Councilmember Gauthier commissioned a study on affordable housing in her district in collaboration with Urban Spatial and the Reinvestment Fund. The report focused greatly on highlighting the socioeconomic profile of the Third district, in addition to exploring changes in housing units over time and the growing problem of housing affordability to the area. However, the report does not go into detail about specific developments, besides analyzing the percentage breakdowns of total housing stock by type (single family, number of units) and highlighting that the “distribution of units by size category has remained relatively consistent since 2011”.
That being said, will there really be enough 20 unit plus projects to make this law impactful?
First, we can use census data to look at year built by property size. The map shows the change in number of 20+ unit buildings between the periods 2000-2009 and 2010-2019.
## Warning: The shape MIN is invalid (after reprojection). See sf::st_is_validThis map shows the difference in units built from 2000 to 2009, and 2010 to present- it does not reflect the change in total units, as it is only focusing on these two periods. We see that the zone in University City does capture one of the tracts that saw the most development of 20+ unit projects in the last 20 years: 24 in the 2000s, and a whopping 258 since 2010. However, the zone for the third district actually sightly misses the mark, only partially covering one tract that has seen 38 of these size developments since 2000. Adjacent to this Mixed Income Zone are two tracts that saw over 100 developments of 20 plus units in the last twenty years, as is seen by the dark green. One could argue that higher numbers of historic developments means there is simply less room for future development of this scale. However, it could also indicate spatial or economic factors that historically discouraged developments of this scale, factors that may persist even after this law is enacted.
That being said, by looking at Census data alone, it seems that City Council did a decent job in drawing these zones. However, total housing units is only one element of the discussion around gentrification and displacement.
Ideally, one could simply identify locations of 20 plus developments using Open Data Philly’s Licensing & Inspections Permit data set, or even possibly the OPA assessor role data, however neither data set offers complete information to discern the development or presence of projects that have at least 20 units and were built in recent years.
However, Open Data Philly does provide a substantial number of building demolition permits.. Demolitions and development go hand in hand- many parts of Philly saw most of their housing stock built in the 1920s and early in the form of rowhomes, so to build a 20+ unit apartment building, there is a good chance demolitions would be involved. In addition, higher number of demolition may generally indicate areas that are drawing consistent development, so these are areas that should be considered for mandatory affordable unit minimums.
We first look at full demolitions in the last three years:
## Warning: The shape MIN is invalid (after reprojection). See sf::st_is_validTable of demolition densities by neighborhood:
It appears that permits are clustering around southwest center city, Kensington, Francisville and Spring Garden. The largest inclusionary zone is in the third district, yet construction permits seem sparse here compared to other regions. In addition, the small zone located in the north, right off Cheltenham, had no full demolition permits in the last three years. This of course comes with normal caveats of public data and reporting, with the possibility of missing or unreported demolitions.
We can draw in census data to understand the socioeconomic context of demolitions. We want to know what type of areas people are demolishing houses in, showing generally where displacement might be taking place. We can explore tracts that have the highest density of permits:
## Warning: The shape MIN is invalid (after reprojection). See sf::st_is_validWe see most demolitions occurring in areas around Sharswood, along with just south of center city. While they aren’t the top tracts in terms of demolition permit density, we do see faint demolition density in the two largest inclusionary zones. The two smallest zones have low to almost no demolition density. This is somewhat concerning; assuming there aren’t massive ammounts of missing data, it would be hard to imagine that these areas would see enough demand for future large scale developments to take advantage of the Mixed Income Overlays if there were little to know demolitions there in the last three years.
However, knowing just where demolitions are happening is only one part of possible analysis into these Mixed Income Overlays. To consider how effective this law is, we want to know what kind of development has already happened in these areas, and what future develop may continue to occur. If we assign each demolition to the median home value and median income of the tract it happened in, we can observe the distribution of home values by count of demolitions.
For reference, in 2019 the median home value in Philadelphia was around $227,00 accoridng to Zillow, and the median income was around $45,00 according the the Census.
## Warning: Removed 12 rows containing non-finite values (stat_bin).
Med income of the tracts demos were located in:
## Warning: Removed 11 rows containing non-finite values (stat_bin).
In both cases, the sample means seem to be around the population means, but both distribution are skewed to the left. It seems that most demolitions are occurring around average valued homes, and in areas with roughly average median incomes. This goes against concerns for gentrification- that most development will be happening in lower income and lower home value areas- but it of course doesn’t say that no gentrification is happening. One could argue that one demolition permit that gentrifies a house in an lower-income area has a greater impact on local socioeconomic conditions than multiple run of the mill home demolitions (demolitions in average neighborhoods to build newer, but similarly valued homes) that presumably populate the data set.
Saving the advanced statistical analysis for economists, we can build a more focused analysis of the socioeconomic context of demolitions. We can pull the top 15 tracts in terms of permit density and look at how their demographics compare to all of Philly.
## Warning: The shape MIN is invalid (after reprojection). See sf::st_is_validIf we relied only on demolition data, City Council’s decision might seem confusing. Only one of the top fifteen tracts in terms of demolition density in the last three years is in a Mixed Income Overlay. Why not make southwest center city a Mixed Income Neighborhood? Most likely, local politics came into play, but also one could consider that the areas of southwest center city with a lot of demolitions are already higher income, so if the point of the law is to protect at risk residents there may not actually be any “at risk” residents in these tracts. However, it still seems notable that the borders drawn by City Council don’t fully capture the areas experiencing the most recent demolitions, and thus arguably the most future interest in development that this law is supposed to be targeting.
We can also turn our analysis to the socioeconomic status of these demolition tracts, to see if maybe the were excluded because they lack the risk of displacement element that seemed to be important in Council’s decision. We can look at the mean value of several demogrpahic factors, comparing the group of just the top fifteen tracts in demolitions, to all of Philadelphia:
| Category | HomeValue | NonWhite | PerCollege | PerBelowPov |
|---|---|---|---|---|
| All Tracts | 202,289.2 | 58.96 | 33.23 | 24.35 |
| Top 15 Tracts | 122,531.1 | 54.00 | 47.00 | 27.00 |
Comparing the two groups, home values are significantly lower in the demolition tracts. However, the top demolition tracts slightly less non-white populations, but significantly higher percentage of population with at least a bachelor’s degree. The percent of population below poverty line was notably lower in the demolition tracts too.
What we are seeing here is that most demolitions are occurring in areas that have lower property values, but have socioeconomic indicators that are equal to, if not slightly better than Philadelphia on average. Notably, there is a higher portion of college educate individuals in these tracts. Pairing this with what we know about gentrification, it seems plausible that the tracts that are seeing the most demolitions are those with desirable property values and local economic factors, which is why we see many demolitions and higher numbers of people with bachelor’s degrees, possibly hinting at gentrifies.
This in mind, maybe our analysis of demolitions alone misses the focus of City Council’s decision- to pick zones and that prevent future displacement. Sadly, maybe these areas have already experienced displacement, as their racial and economic make ups are not exceptionally below the Philadelphia level averages. So, inclusionary zoning might still be effective in creating new affordable units, but it seems that areas with the most demolitions may not necessarily be areas with the most people at risk for displacement.
It really is impossible to talk about gentrification without talking about displacement. It is a challenge to measure displacement, as many data sources require a certain level of anonymity that makes it hard to identify useful details about in and our migration. We can use the Census to look at the percentage of people who lived in the same house one year ago, however:
## Warning: The shape MIN is invalid (after reprojection). See sf::st_is_validHere, we can see a little more why these Overlays could have been drawn. Especially in the third district, we see tracts that have some of the lowest percent of people that lived in the same house a year ago in all of Philadelphia. In other words, these areas have amounts of people who moved in the last year, possibly from a different neighborhood or part of the city. We would expect to see this in both gentrifying areas, and in areas where displaced people are ending up. In either case, an area where more new residents are moving to would be an ideal place to impose inclusionary zoning policies. We do see lower rates in the seventh district, however the two smallest polygons are still baffling. There must be unique example of local development or local political issues that are impossible to pick up via Census or city data that explain their inclusion.
Overall, it seems that City Council did an adequate job drawing these boundaries. While it is concerning that the demolition permit densities were low in these areas, we examine the socioeconomic patterns behind demolitions to see that higher demolition density may not always suggest happen in areas with higher at risk residents. Councilmember Gauthier’s study was significant enough in of itself to show that the largest Mixed Income Overlay is experiencing increased rent burden and possible gentrification, however rationale for the other Overlays is still unclear. Are these areas where developers have expressed interest in moving forward with larger projects, something that our data sets would not be able to pick up? Or are they areas that have the loudest and most active community groups? Either way, when analyzing policies such as the Mixed Income Overlay bills, it is important to consider what factors will allow the policy to achieve its intended result, and also what social and political factors may have influenced their inception.