Intersection of race, redlining, and freeways in Columbus, Ohio: Part 1

 

From my previous post, you know that government of all levels was heavily involved in the systematic discrimination against African Americans, and their short-sighted decisions contributed to a highly segregated America. Black families were refused "bailouts" of their mortgages when the Great Depression hit due to the HOLC's "Residential Security Maps", which were drawn largely on racial lines. Black neighborhoods were redlined, cutting them off of this opportunity for more favorable mortgages, if they were luck enough to own a home in the first place. Racially restrictive covenants, enforced by governments, kept Blacks out of White (or desired to be White) neighborhoods. FDR's Federal Housing Administration would not provide insurance to Black families or developers who desired to build integrated neighborhoods. When other means of segregating the country failed, governments turned to highways, the construction of which displaced countless African American families, without relocation assistance, because they were seen as an "eyesore" to White commuters.

Of course, these sorts of things happened all across the country, and your community is likely no exception. I have been acquiring lots of skills lately in GIS, which is a framework for analyzing geographic data. Using these new skills, I took a dive into the issues of redlining and highway construction as they apply to Columbus, Ohio. I am a big stats guy, so I will look at the numbers and try to identify any sort of patterns that occurred that make discriminatory intent a clear motivation. 

Redlining and Highways in Columbus- Exploration

Using data from the website Mapping Inequality, I created the map below, which is a digitized version of the original HOLC Residential Security Map for Columbus. As you can see, many of the areas near downtown were often colored in red, signifying the occurrences of slum-like living conditions. Recall, however, that areas that were redlined were often done so because they had high Black populations (Rothstein, 2017). Therefore, the areas that were redlined in Columbus can often be considered as those with a high Black population, at least relative to other areas. Present day planning commission names are added to the map to provide spatial orientation.

 

This is pretty much in line with what Rothstein described in his book "The Color of Law". Whites, with substantial backing from the Federal Government, were able to move out to the suburbs. Since banks would not loan to Blacks due to the FHA policy of denying mortgage insurance to lenders who lent to African Americans, Blacks were left behind. Additionally, in the areas where they were allowed to live, African Americans often payed more for similar housing than their White counterparts. With less money for maintenance, areas in city centers often became run down, as clearly shown by the concentration of "redlined" areas around downtown Columbus (Rothstein, 2017). 

Also, recall that interstate highways were often used as a mechanism to both clear Black families out of city centers, a they would be an eyesore to White commuters, as well as reinforcing segregation when other methods, such as racial zoning or racially restrictive covenants, became outlawed by the Supreme Court (Rothstein, 2017).

Since highway planners were encouraged to evaluate the extent to which highways could be used for slum removal (remember, this was a common euphemism for Black), looking at areas that were redlined provide insight, because these were poor, run down communities suffering from disinvestment, and often had high Black populations (Rothstein, 2017). 

Below are a series of maps I created that explore the extent to which highways tore apart redlined areas. Map 2 is the same as map 1 above, except the current freeway system in Columbus is added. Map 3 is similar to map 2, except it shows only HOLC areas that have highways currently running through them. Map 4 is again similar to maps 2 and 3, except in only includes redlined areas in Columbus, regardless if they contained a highway or not. 

Map 2: The digitized HOLC Residential Security Map with the modern freeway system. At a glance, it is clear that areas colored in red were often torn apart by highways.

Map 3: The same as map 2, except it only includes the HOLC areas that had highways either touching them or going through them. Notice there are no "green" areas, and a majority are "red" .

Similar to the other maps, except it only shows areas in Columbus redlined by the HOLC, regardless of whether they had highways through them

At a first glance, it is clear that modern freeway system closely follows those areas that were redlined by the HOLC. Again, these areas were redlined because they were undesirable, with undesirable often meaning "Black". The spatial correlation between redlined areas and freeways is certainly telling, but let's look at some of the numbers. 

Looking at the numbers

Communities torn apart by highways

Of the areas redlined by the HOLC, 53.33% of them had highways tearing them apart. Of those given the grade "C",  22.22% had highways barreling through them. Of areas receiving the grade "B", 12.5% had highways running through them. Finally, there were no areas labeled "A" that had highways running through them. These results are shown in the table below.  

Furthermore, of the HOLC areas through which highways barreled, 40% of them were redlined, 40% were given grade "C", and 20% were given grade "B". Again, 0% of the areas with a highway received the grade "A". These results are shown in the table below.  

These results are compelling, as they suggest redlined areas were more predisposed to be torn apart by freeways. This certainly supports Rothstein's arguments that slums were targeted as new freeway routes.

Adding in racial context

It is important to include racial context to these arguments. To do this, I sifted through the 1950 US Census (which is a massive print document!) to find demographic characteristics for every census tract in Franklin County, including median income and the percent of residents who were Black. Then, using data from the National Historical Geographic Information System (NHGIS), I was able to locate these census tracts spatially and add in the demographic data.

The HOLC data did not include demographic information, so I essentially took the average income and percent Black population for census tracts that are contained in the HOLC areas. So, if a HOLC area contained areas belonging to three different census tracts, the average of those three were used as the single value for the HOLC area. This is not a perfect solution, but it should be pretty close, especially if census tracts near each other have similar demographics, which, given the efforts of governments to segregate America, seems reasonable. This was the best I could do. 

As seen in the table below, redlined areas were on average more Black than areas given other grades. Although the number is not terribly high, it suggests that highways had a disparate impact on Black communities. There are certainly areas that had fairly low Black populations that were redlined, but the numbers suggest that areas that were more Black, at least relative to other areas, were more likely to be redlined, and therefore, more likely to have their lives disrupted by the freeways.

Indeed, a simple logistic regression equation shows that areas that had a 10 percentage point higher Black population were about 21.6 times more likely to be redlined. This supports the idea that areas were often redlined simply because they had high Black populations.  These results are simply correlations, and likely suffer from omitted variable bias, but they are compelling nonetheless. Correlation doesn't prove causation, but it doesn't disprove it either, as for one thing to cause the other, the two things must be correlated. I'll show the regression results below- just so you know I'm not making it up. If you really are interested in the stats, let me know!

The maps below shows the distribution of the percent Black variable for all of Columbus, the areas torn apart by highways, and the redlined areas. The highways are added on top as well. 

This map shows all HOLC areas. It is clear that White neighborhoods were almost always spared from freeways

These are the HOLC areas that had highways running through them. In conjunction with the previous map, it is clear that the areas in the city with the highest Black population very frequently had highways running through them

From these maps, it becomes clear that areas that were the most Black were more likely to have highways running through them, regardless of if they were redlined or not. Indeed, the two HOLC areas that had the highest percentage of Black residents have a highway, or multiple highways, tearing through them. For redlined areas, while there were some with highways that did not have a high percentage of Black residents, it is feasible these areas contained other populations the government wished to oppress- Hispanic, Asian, etc. The point is that areas that were torn apart by highways were, for one reason or another, seen as "undesirable" by the government. Areas that were highly "Non Black" were overwhelmingly spared of freeway destruction.

Additionally, there is a bit of temporal mismatch with the HOLC maps and the 1950 census data. The HOLC Map is from 1936. It is feasible that an area in 1936 was "Definitely declining", but didn't have quite as high of a Black population in 1936 as they did in 1950. Regardless, the maps above show that both redlined areas and areas that were not redlined in 1936, but had high Black populations in 1950, were more likely to be torn apart by a highway.

Another simple logistic regression suggests that areas with a 10 percentage point higher percent Black population are 19 times more likely to have a highway running through their community. Again, this estimate suffers from omitted variable bias, but it's still compelling nonetheless.

Conclusion

In summary, areas that were redlined seem more likely to have been ripped apart by highways. Additionally, areas with higher Black populations appear to be more likely to have been redlined. Therefore, the link between race and highway placement can be clearly seen. Areas that had a high Black population were more likely to be disrupted by highways, regardless of whether they were redlined or not. This is abundantly clear when one considers that higher grade areas (A and B), were almost always spared of this disruption.

The above analysis provides compelling evidence that highway placement in Columbus was made in such a way that had a disparate impact on Black communities. It is very hard to prove discriminatory intent in general, and this analysis is no different. However, I firmly believe that disparate impact is just as bad as discriminatory intent, and while I cannot prove intent, I think the maps and numbers here are convincing in supporting disparate impact. 

Needless to say, Columbus is no different than the stories told by Archer (2020) and Rothstein (2017) about Miami, Pittsburgh, Syracuse, and St. Paul. The racial components of highways are clear, and they separation they perpetuate will be entrenched in our society for years to come.

In an upcoming article, I will do a few case studies of neighborhoods in Columbus, including aerial before and after photographs that show how these White men's roads destroyed Black lives. 

   

Sources:

Archer, D. N. (2021). Transportation Policy and the Underdevelopment of Black Communities. Iowa Law Review, 106(5), 2125-2152.

City of Columbus (2022). Community Planning Areas [data set]. Retrieved from https://opendata.columbus.gov/datasets/community-planning-areas/explore

Steven Manson, Jonathan Schroeder, David Van Riper, Tracy Kugler, and Steven Ruggles. IPUMS National Historical Geographic Information System: Version 16.0 [dataset]. Minneapolis, MN: IPUMS. 2021. http://doi.org/10.18128/D050.V16.0

Mapping Inequality. Columbus Redline Map [data file]. Retrieved from https://dsl.richmond.edu/panorama/redlining/#loc=12/40.01/-82.991&city=columbus-oh&area=B12&text=downloads

Rothstein, R. (2017). The color of law: A forgotten history of how our government segregated America. Liveright. 

United States Census Bureau (1950). 1950 Census of Population: Volume 3. Census Tract Statistics: Part 1. Retrieved from https://www.census.gov/library/publications/1953/dec/population-vol-03.html

United States Census Bureau (2021). TIGER/Line Shapefile for Ohio [data set]. Retrieved from https://www.census.gov/cgi-bin/geo/shapefiles/index.php