The dysfunctional geography of ancient borders in a modern pandemic

Policy driven by municipal and state boundaries makes little sense in COVID-19 fight

LOOKING AT AN aerial photograph or walking down a side street, you’d be hard-pressed to find the border between Waltham’s South Side neighborhood and West Newton. The residential areas on both sides of the municipal line were developed around the same time, in the 1910s and 20s, and the street grid totally ignores the boundary. In fact, there are more than a dozen houses which straddle the two towns: in one of these houses, you might wake up in your bedroom in Newton and then head downstairs for breakfast in the kitchen in Waltham.  

The Waltham-Newton border slicing through streets and houses. (Map by Garrett Dash Nelson, Data: MassGIS)

But when you pass, almost without noticing, from Newton to Waltham, you will have crossed a threshold that makes a lot of difference to public health administrators. As of the middle of November, Newton is in the low-risk “green” category with 9.1 COVID-19 cases per 100,000 residents, while Waltham is a medium-risk “yellow” city with more than double that case rate.  

That’s hardly the only example where being on one side of a border or the other might put you in a sharply different territory of risk. From Brookline, walk south on Russett Road near the VFW Parkway and you’ll cross into Boston—the landscape of suburban single-family homes hardly changes, but you’re now in a community with triple the case rate. Residents of a condominium complex on the southern edge of Springfield pay taxes, vote, and report their health statistics in this high-risk city, but their own swimming pool is located in medium-risk Longmeadow.  

Does crossing one of these invisible lines change your vulnerability to catching the coronavirus? If you’re standing in high-risk Leominster geared up in a mask and face shield, and then take a wide stride over the border into Princeton, which has so few cases that it remains in the state’s “gray” category, can you shed your protective equipment and breathe in deep? Of course not: Political borders don’t matter at all to the virus, and towns haven’t yet enclosed themselves in plexiglass shields. From a purely statistical perspective, it’s ridiculous to pretend that municipalities in Massachusetts make an apples-to-apples comparison with one another.  

The state’s smallest mainland municipality, Monroe, has only 121 residents. It would take a dozen Monroes just to equal the population of a single medium-sized housing development in Boston. Larger cities can conceal striking disparities within their own borders. For example, Malden’s West End neighborhood has household incomes that are much more similar to its wealthier neighbors Melrose and Medford than to the rest of Malden itself. More importantly, the actual geography of how people live and move around has little to do with the boundaries that were laid down, in many cases, by English grantors in the 17th century.  

The chances that one family’s home, grocery store, places of employment, doctor’s office, and outdoor recreation venues all lie within the borders of a single municipality are pretty slim, since we no longer live in the localized agricultural economy that dominated when these borders were drawn centuries ago, and we no longer get around on foot or hoof.  

This is true not just for town boundaries but for state lines as well. For much of the pandemic, Rhode Island has been on Massachusetts’s travel ban. But the statewide case rate in Rhode Island is no higher than the local case rate in a handful of Massachusetts cities. And why prohibit external travel from Rhode Island, which is functionally integrated into the economy and society of the South Coast of Massachusetts, but not internal travel from Berkshire County, which is far more likely to have connections with hard-hit New York City? (In the most recent available data, 65,000 Rhode Island residents commuted into Massachusetts.)  

The only good answer is because there is already a line on the map between Rhode Island and Massachusetts. That’s why we have constructed a pandemic response which is responding as much to the geography of colonial settlement as it is to the geography of this very 21st-century disease.  

Commuter flows in central New England. Boston is far more functionally and economically integrated with Rhode Island and southeastern New Hampshire than it is with western Massachusetts. (Map by Garrett Dash Nelson and Alasdair Rae)

Beginning in August, the state gave regular updates on a “traffic light” map of municipalities, creating a powerful visual reinforcement for the misguided assumption that each town rises and falls as a whole. Recently, recognizing the shortcomings of this cartographic exercise, which over-emphasized risk in some places and under-emphasized it elsewhere, the state has retired this map, choosing instead to provide updates through lengthy reports and tables. But the units in which “high risk communities” are reported have still not changed. Here, it remains every town for itself.  

There are other forms of cartography that could better capture the real spatial distribution of pandemic cases. For instance, a heat map, which aggregates points of infected individuals into “hot” and “cold” areas that ignore administrative boundaries, would offer a more effective method of showing where clusters of cases are occurring. Maps based on specific points must take special care not to violate sick individuals’ privacy rights, by using techniques that add some randomness into location data. And a heat map based solely on home addresses would fail to capture the problem of mobility, since an infected person’s zone of contagion is not actually a single point but rather a line traced along their trips to work, stores, and social engagements. But even with these caveats, a point-based heat map would still offer a much more appropriate visual metaphor for the pandemic’s geography than a series of bounded administrative units that flicker between a handful of solid colors.  

A contrived example of different forms of risk mapping. The left map shows the border between four towns, with points for cases. The middle map is classified by town units, making parts of the green town seem less risky than they really are, and parts of the red town seem more risky than they really are. The map on the right, a heat map, better captures the actual geography of risk. (Maps by Garrett Dash Nelson)

More sophisticated mapmaking, however, cannot solve the real problem of municipal borders, because this is not just a problem about informing the public about which places are risky and which are safe. It’s a problem about how we define the geographic area of responsibility of a public health crisis. The bankers and lawyers who live in low-risk Wellesley have their takeout meals brought to them by drivers who live in high-risk Framingham and send their children to daycare centers staffed by residents of high-risk Brockton. Yet when these service employees get sick, it isn’t the well-resourced health department of Wellesley that tallies them towards its case count or shoulders the burden of ensuring that the infected patients are staying at home. Instead, it’s Framingham and Brockton that have to do these things, and it is these towns that will be ordered to reverse their reopening plans, that will have to close down their schools, and that will begin to fret about the stresses on a fragile municipal budgets 

The pandemic, then, heightens a tension which has been present in Massachusetts since at least the late 19th century, which is the tension caused by a mismatch between political borders and the actual regions in which people—and problems—circulate.  

Some of the first attempts to carve out new jurisdictions for public administration were actually prompted by health concerns. In the 19th century, “sanitary districts,” developed in response to epidemics of waterborne diseases like cholera, allowed early health officials to work in geographies that matched the scale of their technical needs, not the scale of Puritan land ownership. The geographic legacies of local government are hard to break; ever since Boston gave up on annexation in the early 20th century, the municipal boundaries of Massachusetts have remained remarkably unchanged.  

State boundaries are even more rigidly inscribed, with entirely separate governing systems prevailing on either side of a state border.  

Yet the functional geography of the world we live in has continued to evolve, leading policymakers to patch the problem by drawing all manner of special-purpose district geographies, like the five Emergency Medical Service regions of Massachusetts. But do you know which EMS district you live in? Probably not. The issue with these specially drawn geographies is that they exert no force over our sense of collective action or responsibility, and so we go on thinking of ourselves as residents of whatever town we happen to live in— disregarding the fact that our actions affect, and are affected by, a community with a much larger geographic reach.  

Thus we are trying to combat this very 20th century virus, itself swept around the world by the interconnections of a globalized economic system, using political units that are better suited for deciding whose cows get to graze on common land. At the very least, the current pandemic should be treated as an interstate regional problem, addressing the way in which the New England states are knit together in systems of functional interdependence. At the smaller scale, officials should do a better job mapping the fine-grained geography of where outbreaks are actually occurring within and across municipal lines.  

Meet the Author
But what is most important is to recognize that our centuries-old geographies simply do not help us very much in understanding—or fighting—the spread of a problem that follows the spatial patterns of the present day. 

Garrett Dash Nelson is curator of maps and director of geographic scholarship at the Norman B. Leventhal Map and Education Center at the Boston Public Library.