Baltimore Eviction Map

By: Tim Thomas, Ian Kennedy, Alex Ramiller, Ott Toomet, & Jose Hernandez
May 8th, 2020


The Eviction Study's Baltimore Map showcases data collected by our team as a dynamic and interactive visualization, highlighting which neighborhoods are at the highest risk of scheduled evictions and eviction removals. We analyzed a sample of 9,349 scheduled evictions between 2018 and 2019 where 66% of these cases led to a physical removal. Comparing this sample to US Census data, we estimate that between July 2018 and June 2019:

  • 46% more female headed households were removed from their homes as compared to male headed households, a ratio of about 3 women for ever 2 men.

  • The number of Black eviction removals is 3 times higher (195% more) than the white eviction count (4,775 Black evictions vs. 1,614 White evictions).

  • The number of Black female headed household removals is 3.9 times higher (296% more) than the number of white male headed evictions (2,996 vs. 775) and 2.3 times higher for Black male headed households.

  • 7.3% of all Black male headed households and 5.4% of all Black female headed households were removed from their homes. These rates are roughly 51% and 11% higher than the White male headed eviction rate.

  • The map shows that the highest risk of eviction occurs in the most segregated neighborhoods to the West and in gentrifying neighborhoods to the East.
Demographic, income, and rent characteristics are also included in the map to provide context and highlight the attributes of places where evictions tend to occur. Our demographic estimates show a massive racial disparity making evictions a civil rights issue related to contemporary discrimination in housing access, displacement, and economic inequality linked to the legacies of segregation, policies, and practices directed against persons of color.

At the time of writing this article, COVID-19 is sweeping the nation and it is unclear how the pandemic will end. However, what is clear, is that housing is vital to mitigating the spread of the virus and sustaining health. The racial disparity of infection is exceptionally concerning as it reflects long-lasting legacies of structural discrimination in housing, health, and wealth. This trend suggests that communities of color will be the hardest hit, deepening disparities, requiring immediate intervention to improve public health.

Note that data are not available for every month in our time frame meaning estimates may slightly change as we add more data.

Baltimore Summary of Findings

Over the past decade, research has shown that evictions reinforce poverty and limits housing opportunities for the nation's most economically vulnerable. The mark of an eviction sets households on a path of housing insecurity that can prevent them from living in better neighborhoods, establishing beneficial networks and economic opportunities, increases health problems, and is one of the leading indicators of future homelessness. With over 150,000 filings per year, Baltimore's eviction rate is one of the highest in the nation, requiring investigation to why this issue persists and policital intervention.

This study examines a sample of 9,349 scheduled evictions provided by the Baltimore City's Sheriff's Department from January, 2018 to July, 2019. During this time period, about 65.7% of all scheduled evictions ended in a physical removal. In each of these scenarios, we find a severe racial disparity in both scheduled and physical evictions.

What this map offers

The Baltimore Eviction Map provides an estimated risk of scheduled evictions and eviction removals for a given neighborhood from our full sample (9,349) between January, 2018 to July, 2019. By sample, we mean that this is not the full count of evictions during this period, but, it is large enough to estimate risks and overall rates. By risks, we mean we divide the number of evictions in a neighborhood by the number of renters in the same neighborhood and then compare that neighborhood rate to the rest of the city. This is the same methodology used by health researchers to estimate risks of infection. Eviction risk ranges from 0 (extremely low) to 4.01 (extremely high).

The map legend provides 7 categorical risk levels. As compared to the rest of the city:

  • extremely high = +3x higher risk
  • high = 2x to 3x higher
  • above normal = 1.1x to 2x higher
  • normal = 0.9x to 1.1x
  • below normal = 1/2 to 9/10 lower risk
  • low = 1/4 to 1/2 lower
  • extremely low = 0 to 1/4 the risk

We also provide map layers of racial segregation and the locations of historical redlining zones. You can also click on any tract to view other neighborhood information such as risk estimates, housing costs, income, housing types, and demographic compositions like race, percent of children, and percent elderly.

Map Trends

Overall, we see that the risk of eviction is highest in the Western Black segregated neighborhoods, gentrifying Northeastern neighborhoods, and racially diverse South Baltimore neighborhoods. Areas with larger White and wealthier compositions have the lowest risks of gentrification.

City Eviction Rate Trends

Using advanced statistical models, we estimate Baltimore's eviction rates and ratios by gender and race using a 38% sample of the 6,519 household eviction removals during the 2019 fiscal year (July 1, 2018 to June 30, 2019) described in the table below. Count ratios compare the number of evictions for one group as compared to men, Whites, White men, and White women. An eviction rate is the percent of the respective population that was removed. Rate ratios compare the respective population eviction rates to the same four categories previously mentioned.

The overall eviction rate for Baltimore is 5.3%, which is 2.3 times higher than the national eviction rate of 2.3%.


Female headed households were evicted 1.46 times more than men, or 46% more than men. However, there are more female headed households in Baltimore so the female headed household eviction rate ratio is 9% lower for women than men (5.1% vs. 5.6%).


Black headed households had the highest removal count which was 2.96 times higher than White removals (4,776 removals vs. 1,614 removals). The Black eviction rate is 5.9% (13% higher than the White eviction rate of 5.2%). Latnix and Asian households had an eviction rate of 2.2% and 0.5%, respectively.

Race and Gender

Finally, Black female headed households had the highest number of eviction removals (2,996, which is 5.4% of the Black female headead households) followed by Black male headed households (1,806 evictions at 7.3% of the Black male headed population), White female headed households (807 at 5.5%), and White Male headed households (775 at 4.8%). Asian, Latinx, and other groups had under 50 removals during the fiscal year. The plot below shows the differences in counts and rates (percent of the population that was removed).

Race Gender Estimated Total Eviction Count Count Ratio to Men Count Ratio to Whites Count Ratio to White Men Count Ratio to White Women Total Renting Households Estimated Eviction Rate Rate Ratio to Men Rate Ratio to Whites Rate Ratio to White Men Rate Ratio to White Women
-- female 3,868 1.46 -- -- -- 75,532 5.1% 0.91 -- -- --
-- male 2,651 1.00 -- -- -- 46,977 5.6% 1.00 -- -- --
Asian -- 21 -- 0.01 -- -- 3,858 0.5% -- 0.10 -- --
Black -- 4,776 -- 2.96 -- -- 80,518 5.9% -- 1.13 -- --
Latinx -- 103 -- 0.06 -- -- 4,680 2.2% -- 0.42 -- --
other -- 5 -- 0.00 -- -- 2,650 0.2% -- 0.04 -- --
White -- 1,614 -- 1.00 -- -- 30,803 5.2% -- 1.00 -- --
Asian female 13 -- -- 0.02 0.02 1,558 0.8% -- -- 0.17 0.15
Asian male 6 -- -- 0.01 0.01 2,300 0.3% -- -- 0.06 0.05
Black female 2,996 -- -- 3.86 3.71 55,887 5.4% -- -- 1.11 0.98
Black male 1,806 -- -- 2.33 2.24 24,631 7.3% -- -- 1.51 1.35
Latinx female 48 -- -- 0.06 0.06 2,037 2.3% -- -- 0.48 0.43
Latinx male 64 -- -- 0.08 0.08 2,643 2.4% -- -- 0.50 0.44
other female 3 -- -- 0.00 0.00 1,240 0.3% -- -- 0.05 0.05
White female 807 -- -- 1.04 1.00 14,810 5.5% -- -- 1.12 1.00
White male 775 -- -- 1.00 0.96 15,993 4.8% -- -- 1.00 0.89

Policy Recommendations

Given the severe racial disparity in evictions, the consequences of losing your home, rapidly rising rents, and the growing economic divide, we believe that several policies should be enacted to protect citizens in Baltimore and abroad. Based on our observations of what works in other states, here is a short list to consider:

  • Provide a right to counsel in eviction cases.
  • Require the landlord to send a pre-filing notice before filing any eviction case.
  • Reform habitability laws to make it easier to for tenants to defend their eviction cases based on conditions of serious disrepair.
  • Prevent eviction records from being used in the rental application process.
  • Prohibit no-cause tenancy terminations and enact rent stabilization, i.e., limit on rent increases.
  • Invest in permanently affordable, community controlled housing and rental assistance where needed.
  • Revise the current inclusionary housing law to produce more units for lower incomes.
  • Remove no-cause terminations.
  • Prevent excessive rent hikes based on tenant income.

Our goal in releasing these maps and reports is to follow in the footsteps of other collective and scholarly researchers and reveal previously unknown patterns of eviction while inspiring more research and discussion on the links of neighborhood change and housing precarity in the United States. We are currently examining other cities and states and will release those reports as they become available. So, please visit us again and follow our progress on twitter.

Data and Methods

This map uses data on a sample of 9,349 scheduled evictions provided by the Baltimore City Sheriff's Department from January 2018 to July 2019. The overall city rates and ratios are calculated using a 37% sample of 6,519 cases that occurred during 2019 fiscal year starting on July 1, 2018 and ending on June 30, 2019.

Tenant race estimates are calculated using a Bayesian prediction model developed by Kosuke Imai and Kabir Khanna using the surname and geolocation of the tenant. These imputed estimates produce some possible uncertainty. However, we compared our estimates to actual intake data from the King County Bar Housing Justice Project (HJP) and find our rates are within a few percentage points.

All other data - including renting population, median household income, median rent, and rent burden - originates from the 5-Year American Community Surveys (ACS) and 1 year 2017 IPUMS dataset conducted by the US Census Bureau. This map displays the five-year estimates for 2014-2018.

Eviction rate and relative risk estimates for the given race are calculated using the estimated number of renting households recorded in the ACS. If less than 100 renting households are recorded within any given census geography, or if the margin of error for the estimated population of renting households is larger than the estimate itself, these estimates are omitted. Eviction rate represents the number of evictions within a given census geography divided by the estimated number of renting households. For any given race, the relative risk of eviction is based on the rate of evictions among renting households within each census geography divided by the rate of evictions across the entire study area.


This interactive map interface was created by Tim Thomas with assistance from Alex Ramiller, Ian Kennedy, Ott Toomet, and Jose Hernandez. The Evictions Study would not be possible without the generous support from Bill Howe and the Cascadia Urban Analytic Cooperative, the University of Washington eScience Institute, the University of Washington Center for Studies in Demography and Ecology, The University of California at Berkeley Center for Community Innovation and the Urban Displacement Project, Enterprise Community Partners, The Public Justice Center in Baltimore, The Baltimore City Sheriff's Department, Fair Development Roundtable, Baltimore Renters United, United Workers, Coppin Heights Community Development Corporation, Charm City Land Trusts, Jews United for Justice, Professor Malcom Drewery at Coppin State University, Professor Meredith Greif at Johns Hopkins University, and Linda Morris at the ACLU. This application was built using the R statistical program and Leaflet.