Posted on December 17, 2019
(Note: This research brief was written by Tom Hopper, MHP Director of Research & Analytics. It is also available at www.mhp.net/todex).
BOSTON, Dec. 18, 2019 --- MHP's Center for Housing Data has developed a new methodology for estimating the number of homes at the parcel/lot level. This new method employs multiple data sets in order to create a consistent, accurate, and comparable metric that can be used to compare housing patterns across municipal borders.
We are releasing a progress report on this effort today through Transit-Oriented Development Explorer (TODEX), a new web site that includes a user-friendly tool that allows you to review densities at all 261 Greater Boston transit stations.
This research brief - slightly revised here and also available on the TODEX web site at www.mhp.net/todex - serves as a progress report on our efforts and initial findings, including a discussion of potential next steps. The methodology we used to develop our housing estimates and the creation of station areas is explained on the TODEX web site. Those who are interested in the details of the analytical process should check out the 'Methodology' tab for details, or reach out to MHP's Tom Hopper or Lucas Munson with questions or feedback. The research is evolving and we would love to hear your thoughts as we refine this work.
We embarked on mapping and analyzing all 261 transit stations because the benefits of transit oriented development (TOD) are well-documented. By encouraging concentrations of housing, jobs, services and recreation in transit-accessible locations, regions can increase transit ridership, which in turn can serve to enhance the viability and cost-effectiveness of transit systems. TOD increases access to work, services, and other destinations by increasing mobility, which is especially important for low-income households. Additionally, an increase in transit usage results in a reduced reliance on single occupancy vehicles, which is positive for both the environment and traffic congestion. Walkable, transit-rich neighborhoods are increasingly sought after across the demographic spectrum, including young professionals and older households hoping to down-size. Creating more of these neighborhoods could serve to reduce cost pressures and competition for the limited number of existing places of this type (perhaps even helping to curb displacement of low-income households).
Massachusetts is simultaneously experiencing transit and housing crises. We have an acute housing shortage, and this has created an incredibly expensive and competitive housing market that provides few opportunities and many obstacles for low and moderate income households. iii Combined with a need for improvements and investments in transit infrastructure and service, as well as some of the worst traffic in the country, these compounding issues threaten the viability of the region and the well-being of residents. Greater Boston needs a comprehensive vision that treats housing and transit issues in concert. Transit-oriented development strategies are at the nexus of these two critical public policy problems.
By encouraging housing, jobs, services and recreation in transit-accessible locations, regions can increase transit ridership, which in turn can serve to enhance the viability of transit systems. This also increases access to work and services, which is especially important for low-income households.
Changing the way we develop around transit stations will take a lot of regional planning. Planning on a regional scale in Massachusetts is a very difficult task, however, especially when it comes to housing and land use. Each of the 351 cities and towns in Massachusetts sets its own zoning and land use regulations, and the state does not impose many zoning requirements on cities and towns. ii Due to this patchwork of land use regulations, there is wide variation in development patterns across neighborhoods served by our transit systems. These variations have not been regularly measured or evaluated, in part because of a lack of availability and consistency in the data available from town to town. This research brief is a progress report on our ongoing effort to collect and analyze parcel data across the Commonwealth. To see the full report, methodology and TODEX tool, go to www.mhp.net/todex.
Findings and observations
Homes are limited in many station areas across Greater Boston.
We took every station on the commuter rail, subway (orange, red, and blue lines) and light rail/bus lines (green and silver lines), constructed a station area around it, and measured the number of homes for every parcel in the area. We then aggregated these housing counts to create an overall estimate of homes per acre.
The below plot shows the distribution of gross density across the 261 station areas in our analysis, grouped by service type. The distribution is heavily skewed toward the lower end of the spectrum. The median across all station areas is roughly 6.2 homes per acre, with most of the highest density areas located along rapid transit lines in the City of Boston and a handful of inner core communities, and most of the lowest density areas located along the commuter rail. Commuter rail station areas had a median of 2.8 homes per acre, rapid transit station areas had a median of 13.1 homes per acre, and light rail/bus (silver, green line, and Mattapan trolley) stations had a median of 15.2 homes per acre.
These differences in housing patterns across transit types are readily apparent in the adjacent column chart, which shows station areas ranked by units per acre and filled based on route type. Neighborhoods near light rail and subway stations have higher levels of residential density than those surrounding commuter rail stations.
There is no single, universal minimum density requirement that can be applied to all locations. Service levels vary along with issues of connectivity and physical barriers to access. In some places, while development may not be supported by the market, transit service may still be an important in order to provide equitable access and mobility to residents in these locations. Still, research on minimum densities needed to support transit service can be instructive as we try to compare the neighborhoods served by transit here in Massachusetts. In a 2011 analysis, Cervero and Guerra concluded that 30-67 residents per gross acre are required to achieve cost-effective light rail investments (the range depends on capital costs). iv If we make a very crude assumption and apply average household size of 2.54 in Massachusetts, v we could interpret this as a 12-26 housing units per acre range. On the lower end of estimates, in the mid-1970s, researchers Pushkarev and Zupan concluded that nine households per acre could support light rail while twelve households per acre could support heavy rail. TOD guidelines from a variety of other regions recommend a minimum of 15 units per acre along arterial corridors, and even higher for town centers and urban core neighborhoods. vi,vii Given these ranges, and since this initial analysis is just a first look, we believe a basic 10 residential units per acre threshold would be useful in evaluating performance across the neighborhoods in our study.
Most station areas in Greater Boston have fewer than ten homes per acre.
Of the 261 station areas in our study, 162 have fewer than 10 homes per acre. This constitutes about 62 percent of all station areas, and approximately 78 percent of the gross total land in our station area data set. Of the station areas with fewer than 10 residential units per acre, 119 have an average of less than five homes per acre. These stations represent 46 percent of station areas and 60 percent of land in station areas. Many of these stations are located along the commuter rail.
What does ten homes per acre look like? These numbers seem abstract, so it is helpful to visualize what this means. The Urbanist did a great piece in 2017 that explores what different levels of density look like, with examples from Washington State. ix The example they use for 9.3 units per acre shows that these concentrations of homes can be achieved even with a mix of triplexes, duplexes and single-family homes. So, when we say 10 units per acre, we do not mean Manhattan. 10 units per acre is a bare minimum that should be attainable in practically any context. In our research, many of the station areas are in primarily residential areas that do not have high rises or large complexes. Great and dynamic neighborhoods in Dorchester and Jamaica Plain are great examples in the City of Boston, and places such as downtown Gloucester and Swampscott (even though they are slightly below ten homes per acre) show that you can approach these density levels outside of the inner core. These are all places that we love, and feel walkable and convenient. Unfortunately, these places are few and far between along our transit corridors. There is clearly room for improvement.
If we were to increase housing opportunities in station areas served by the commuter rail and rapid transit to reasonable levels, we could generate approximately 250,000 additional housing units.
Our analysis created 261 distinct station areas, totaling 72,197 gross acres in aggregate. Across all of these areas, we estimate there are approximately 469,000 housing units, which yields an average of about 6.5 homes per gross acre across the system. If we were to achieve a modest increase to ten homes per acre across these neighborhoods, we could have approximately 253,000 additional housing units, all in locations with good access to transit infrastructure. As the data improves and we have better information on more recent developments this overall number might shift slightly, but we believe this is a fair estimate of the potential these key neighborhoods have in providing homes for Massachusetts residents.
While this math is incredibly simple and ignores some important neighborhood factors, it does show the potential that reimagining these high-access neighborhoods could have in terms of better supporting transit while simultaneously making a huge dent in our chronic housing supply problem. Of course, every station area is different, and there are places where even greater levels of density make sense, and surely some places where market demand may not support large numbers of additional new units. Still, neighborhoods that are connected to transit infrastructure represent our best chance for developing in a more sustainable way.
The highest and lowest density station areas demonstrate the range of land use intensity we are working with, ranging from 0.02 units per acre to 80 units per acre.
The highest density station area is a 45 acre neighborhood surrounding Symphony station in Boston, served by the green line. This small area contains 3,622 homes, which results in roughly 80 homes per acre. In contrast, the lowest density can be found on the 474 acre area surrounding the Grafton commuter rail station, served by the Worcester Line. There are only 10 homes in this area, resulting in just 0.02 homes per acre. While there are a handful of other uses near the station, housing is clearly not part of the development pattern.
Density does exist on the commuter rail, but urban areas are the only places in which the system is meeting reasonable levels.
As one might expect, most of the densest station areas on the commuter rail are within the City of Boston, including Landsdowne station in the Fenway-Kenmore area (23.7 units per acre) and several stops along the Fairmount Line in Roxbury and Mattapan (Four Corners, Uphams Corner, and Talbot Ave at 15.1, 14.4 and 12.2 units per acre, respectively). Along the Fairmount corridor, most residents do not have access to a rapid transit station within walking distance but do have good access to bus networks. Rail service in these areas, especially if frequency increases in the future, is an amazing asset to these communities that have housing levels needed to support that service.
The most dense station area on the commuter rail that is outside the City of Boston is Haverhill station, with an average density of 11.8 units per acre, followed by Lynn station at 11.5 units per acre, Beverly station at 11.3 units per acre, and Waltham station at 11.2 units per acre. These urban locations demonstrate that higher levels of density outside the immediate urban core are attainable, and that the commuter rail helps to connect these hubs via transit.
Mixed use neighborhoods on rapid transit and light rail lines stand out as good examples of transit-oriented development.
The highest density locations on rapid transit and light rail tend to have a mix of uses represented near stations. While our analysis has not yet been extended to non-residential properties, most gray parcels on our map represent non-housing uses and help provide a broad sense of how much land is dedicated to these other purposes. Places such as Porter Square, Kenmore Square, Malden Center, and JFK/UMass are all good examples of mixed-use neighborhoods.
Some suburban station areas represent squandered opportunities, while others might still offer opportunities for new housing.
The least dense station areas are a mixed bag. There are some station areas that are almost entirely low-density residential, predominantly single family homes on large lots. These station areas tend to have few non-residential properties. Weston and Wellesley each have three stations, and all of these station areas are prime examples of this pattern. Wellesley Farms and Silver Hill stations, pictured below, have gross density levels of 1.6 and 0.8 units per acre, respectively. How do we make better use of our transit investments in these neighborhoods when they are already built out at extremely low density levels?
In contrast, there are other station areas that have little housing but that are also very dense commercial areas. This is certainly the case for Kendall Square in Cambridge, where land is undeniably used intensively, just not for housing. While Kendall Square is thriving, we should be asking if housing is an important, but neglected component of neighborhoods like these. Is there an opportunity to turn neighborhoods like Kendall Square into one where people not only work, but live? The neighborhood around Newmarket station in South Boston/Dorchester is another location in which there are a number of jobs and commercial space (particularly around the South Bay shopping center). The data from Boston is from 2017, and we are aware of a number of new developments in this area that may not yet be captured in our analysis. As this neighborhood develops and becomes better connected with residential uses, it has the potential to develop into a successful mixed use neighborhood.
Differences in density can be observed in neighboring communities with the same level of service, suggesting that municipal zoning and land use planning is driving much of the variation we see.
There are inconsistencies across locations that receive the same level of service. The two branches of the red line are a prime example of this. While the four stations of the Ashmont branch in Dorchester all have average densities of 10 units per acre or higher, the five stations along the Braintree line have significantly lower average residential densities that range from just 2.6 to 6.4 units per acre. While ridership may be high on the Braintree line, many commuters accessing these stations are likely driving from locations south of Boston. The stations along the Ashmont line are more likely drawing riders who live in the area. Further research on ridership, parking, commuting patterns and habits might shed more light on this.
We have only begun to scratch the surface in terms of comparisons that might tell compelling stories about the variation we are seeing across these neighborhoods. Certainly there are stories to be told about equity, zoning, and quality of transit services through some of these images. We would love to hear from our readers about what they are seeing as they explore these maps, and whether their experience living in or traveling through these neighborhoods is reflected in the visualizations we've created.
Policy discussion and future research
This first glimpse at residential density makes it clear that there is ample variation in the built environment around the Greater Boston transit system. This variation is reason enough for transit and housing policy professionals to think jointly about transit access and residential development patterns when considering service enhancements or extension to new areas. To the extent that local action is squandering the opportunities that could accrue from transit access, bold action may be necessary. There should be an expectation that investments in mobility are met with a commitment to not only allow, but to thoughtfully foster the kind of dense, walkable neighborhoods that yield the desired benefits. Our inability to make this connection will continue to hamper our ability to tackle our transit and housing crises.
We are encouraged by these initial results from a methodological perspective, but there are clearly some additional data points that would help to enhance our understanding of what is really happening in station areas. The utility of this data set would be greatly enhanced by combining it with point-level metrics on job locations, parking availability and requirements, and information on local zoning around transit stations.
There are several additional data points that could be beneficial for future research. First, critical density thresholds often use a combination of residents, students and jobs. x,xi Our research created a methodology for measuring the number of housing units in station areas, but does not take the next step of extrapolating a population estimate from those unit counts, nor does it incorporate data on the number of jobs in these areas. Further research should attempt to paint a more accurate picture of these station areas by incorporating these other factors. Research has even shown that adding jobs to a station area has a greater impact on ridership than adding residents. xii
Second, this initial research focused on rail (and the silver line). Our methodology should be extended to include areas served by major bus routes, since many of the inner ring suburbs have express buses that are heavily used. The availability of multiple modes of transportation is an enormous benefit to a neighborhood, and likely impacts commuting decisions and could even warrant higher levels of density.
Last, the methods developed in this paper should be further refined. Station areas were determined by drawing a half mile buffer around each station using Euclidean distance, which is a crude approach. Road networks could be used to create a service network that better maps out level of access via a half-mile of roadway. While this methodological paper established that residential density could be measured in transit-rich areas and that there is significant variation in development intensity across the system, there is much work to be done to create a clearer picture of how stations are accessed and how people make decisions around which mode of transportation to take to reach their destinations.
This work becomes better with the help of your insight and feedback. Please reach out to Tom Hopper, Director of Research & Analytics with questions, comments, corrections and suggestions.
i Metropolitan Area Planning Council. 2012. 'Growing Station Areas: The Variety and Potential of Transit Oriented Development in Metro Boston.'
ii Modestino, Alicia, Clark Ziegler, Tom Hopper, Calandra Clark, Mark Melnik, Carrie Bernstein, Lucas Munson, and Abby Raisz. 2019. Greater Boston Housing Report Card 2019. Boston: The Boston Foundation.
iii Modestino, et al.
iv Cervero, Robert, and Erick Guerra. 2011. Urban Densities and Transit: A Multi-dimensional Perspective. Working paper, UC Berkeley Center for Future Urban Transport.
v U.S. Census Bureau; American Community Survey, 2017 American Community Survey 1-Year Estimates
vi Boris Pushkarev and Jeffrey Zupan. 1977. Public Transportation and Land Use Policy, Bloomington: Indiana University Press.
vii MARTA (Metropolitan Atlanta Rapid Transit Authority). 2010. 'Transit-Oriented Development Guidelines.' Atlanta.https://www.itsmarta.com/uploa...
viii Metropolitan Council. 2006. 'Land-Use Densities: Rules of Thumb.' https://metrocouncil.org/Communities/Services/Livable-Communities-Grants/Maps,-forms-misc/Metropolitan-Council-TOD-Guide-Land-Use-Densities.aspx.
ix Bengford, Bob. 'Visualizing Compatible Density.' The Urbanist. May 2017. https://www.theurbanist.org/20...
x Puget Sound Regional Council. 2015. 'Transit-Supportive Densities and Land Uses.' Seattle.
xi Cervero, Robert, and Erick Guerra. 2011. Urban Densities and Transit: A Multi-dimensional Perspective. Working paper, UC Berkeley Center for Future Urban Transport.'
xii Puget Sound Regional Council (16).