By David Han, Tzion Jones, Zoey Katzive, and Yicheng Lu

In this exploratory analysis, we outline potential predictors of housing insecurity with a focus on details relevant to those falling through the cracks of our housing system.

Why this matters:

Housing insecurity is a particularly important issue because it greatly influences the educational outcomes, health, and economic stability of affected individuals. While the effects of Rhode Island’s ongoing housing crisis can be felt by all homeowners and renters, this burden is distributed disproportionally across the state. Traditional measures of homelessness or housing stock shortages may not capture the needs of people on the brink of disaster, which could be a key group to study to understand how the housing crisis can be addressed effectively. 

What can be done:

In compiling this data we hope to provide a resource to assistance organizations that are interested in helping those affected but often struggle to reach people before they are already facing an eviction or other housing emergency. As it stands, these organizations often use indirect metrics as proxy estimates for the size and spread of their audience. By illustrating the geographic areas most  vulnerable to either eviction or our identified precursors, we seek to focus these organizations’ efforts, provide a starting point for further research, and to inform prospective renters and residents of apparent trends in their area. 

Figure 1. Map of eviction rate by zip code in Rhode Island.

While high eviction rates are a problem across the state, urban Rhode Islanders, as well as those in northeastern zip codes may be at greater risk. Between 2021 and 2023, the zip code with the highest eviction rate was 02903, comprising much of Providence and totaling approximately 14.2 evictions per 100 residents. Other zip codes surrounding Rhode Island’s largest cities such as Cranston, Pawtucket, and Woonsocket also experienced disproportionately high rates of eviction. Given that eviction rates vary geographically within the state, it is important to consider why that might be the case. 

In the next sections, we will examine the relationship between eviction and three socioeconomic factors. Namely, we explore correlative trends in the income of renters, the affordability of housing, and the housing infrastructure itself.

Eviction, Income, and Housing Costs

The increase in utility costs has also been sharp, with Desmond reporting a 53% increase from 2000 to date. Paichen Li (2021) finds a positive correlation between average energy burden and rate of eviction within a given census tract, further suggesting that poverty, and ostensibly housing costs, remain a driving factor of the crisis.  It stands to reason that the housing crisis will disproportionately affect those low-income households already struggling to make ends meet. 

Turning briefly to potential social determinants of household income, Seungbom Kang (2019) finds family employment structure, job insecurity, number of adult family members and number of children to be important household-level predictors of housing insecurity, controlling for both income and housing cost. 

Households with adults suffering from serious health issues also experience housing insecurity more often than those without, but the literature indicates this may have more to do with pre-existing differences between households than health status itself. Findings by Lind & Egede (2023) suggest that expansion of Medicaid can reduce both the eviction rate and the overall number of eviction filings, further supporting the role of health expenses in housing affordability.¹² It may be that the prospect of sudden medical bills prevents households from safely committing their income to rent and utilities beyond a certain threshold (e.g. 30%), leaving them more susceptible to eviction in hard times. This would align well with Kang’s finding that low-income households are at higher risk of eviction when sources of income or employment are unstable, since medical needs can add a degree of uncertainty to one’s regular expenses. 

The general correlation between eviction, housing affordability, and lower income is reflected in our own research, as the following series of figures will demonstrate. We find a significant correlation between the median household income and eviction rate across Rhode Island zip codes. 

Figure 2. Scatter plot showing the relationship between median household income and evictions per 100 residents, for each zip code in Rhode Island. The data is represented by points for each zip code and a line of best fit.

Table 1: Median household income, eviction rate, and population for the 10 lowest-income zip codes in Rhode Island.

Figure 3. Difference in average rental costs (2-bedroom unit, including utilities) in Rhode Island municipalities from 2018 to 2023. All values in U.S. dollars. Positive values to the right indicate rent increases and vice versa. For instance, West Greenwich leads with the highest rent increase of $1,460.17, in stark contrast to Westerly, which shows the most significant decrease of $403.42.

Figure 3 plots the average rental cost changes across Rhode Island municipalities from 2018 to 2023, based on data from the RIHousing Rent Survey. Displayed in 2023 dollars for consistency, the data combines average rent figures for 2-bedroom units from CoStar with estimated utility costs from the 1-year American Community Survey (ACS). The graph omits municipalities with fewer than four sample properties to maintain data reliability. Municipalities are ranked by the extent of change, showcasing areas with the greatest increases down to those with the most significant decreases in rental costs.

Figure 4. Line graph tracking eviction filings and utility prices in Rhode Island from January 2020 to January 2024. The left y-axis shows the number of Eviction filings. Utility prices, including gas (measured in dollars per thousand cubic feet, MCF) and electricity (in cents per kWh), are shown with the right y-axis. The data points are shown with different colors and styles. Interconnected by lines, the graph depicts the trend over the 4-years period.

Note: Eviction filings data for Figure 4 are sourced from The Eviction Lab at Princeton University. Residential natural gas prices, measured in dollars per MCF, and electricity prices, measured in cents per kWh, are sourced from the U.S. Energy Information Administration. Geographic information is aligned using the Zip/Tract Bridge File provided by the U.S. Department of Housing.

Eviction and the Age of Housing Stock

Figure 5. Scatter plot representing the relationship between pre-war units per 100 residents and evictions per 100 residents in Rhode Island. Each point represents a unique zip code. 

Note: The “pre-war units per 100 residents” variable was calculated by dividing each zip code’s total count of units built before 1939, dividing by population, and multiplying that number by 100. 

The literature also suggests that the quality of housing units themselves can be a determinant of housing insecurity. In an analysis of 71 metropolitan areas across the United States, Hepburn et al. found median eviction rates to be significantly higher in older urban neighborhoods with housing stock constructed primarily before 1970.1 One explanation from the Harvard Joint Center for Housing Studies posits that the cost of improvement and repair runs much higher for old housing units as their core components grow worn or obsolete.13 For low-income households, maintenance of utilities and other important systems may represent a larger portion of their overall income, and create larger burdens.  

That in mind, our final set of observations concern seasonal fluctuations in eviction filing, as well as general trends in the frequency of eviction over time.

Figure 6. Number of eviction filings by month in Rhode Island from 2020 to 2023.

Figure 6 shows a year over year increase in the number of eviction filings in Rhode Island. There was a 9.9% increase in filings from 2020 to 2021, a 24.2% increase in filings from 2021 to 2022, and a 16.9% increase in filings from 2022 to 2023. According to a 2023 report from the RI Department of Housing, housing unit production in the state dropped to under 1000 permits per year after the 2008 crisis and has continued to remain low.7 The trend in eviction filings matches a trend of decreasing year over year housing unit production for the first time since 2011. The shortage of affordable housing in Rhode Island may be a byproduct of housing unit production not meeting the supply of people looking to rent.

We can understand seasonal variations in eviction by aggregating the number of filings occurring in each month. We observed that the highest amount of filings occurred in January and the period from August to October. The lowest amount of filings occurred in April and May. The Eviction Lab described similar variations in evictions across the whole country as a predictable seasonal pattern.6 Our findings follow trends in the housing market found by the Bureau of Labor Statistics (BLS) on the starting dates of leases by month8. The highest proportion of starting dates were in the months of June to August, and the lowest proportion was found from December to February. These correlations suggest that evictions occur most frequently during the same times that people are looking for housing, in the summer and fall months.