Metros with Faster High-Income Renter Growth Saw Larger Cost Burden Increases

Key Takeaways

Since 2012, metros that added high-income renters fastest saw larger cost burden increases for renters lower down the income distribution.

The pattern is specific to high-income demand, as when metros are grouped on middle- or low-income renter growth, there were not comparable gaps in cost burden increases.

Cost burdens for lower-income renters accelerated in the metros that joined the fast-growth group after the pandemic and kept rising in the metros where high-income growth faded.

Introduction

Renter affordability has not eroded evenly, and household formation trends help illustrate why.

High-income households, defined here as those making over $75,000 in 2010 dollars, emerged from the Great Financial Crisis (GFC) as the fastest-growing renter cohort, growing 60.7% between 2010 and 2019 compared to 11.5% for middle-income renters (those making between $30,000 and $75,000[1]) and -3.3% for low-income renters (those making less than $30,000)

Rental household growth has been more balanced post-pandemic, yet during both periods, cost burdens for low- and middle-income renters accelerated while remaining tepid for high-income renters.

On the surface, this is explainable—households with higher incomes are more likely to rent by choice, often accepting higher rents without new constraints on their incomes. However, in a recent Chandan Economics analysis of the national K-shaped rental affordability crisis, we hypothesized that a type of cost-burden filtering could be at play, where, as high-income renter growth accelerates, increased cost pressures show up down the income stream.

To scrutinize this hypothesis, we designed a metro-level analysis that more directly tests whether cost-burden filtering is occurring. If cost burden filtering is happening, metros that add high-income renters the fastest should, relative to other metros, see larger cost burden increases for renters lower down the income distribution over the same period. Ranking metros by low- and middle-income renter growth should produce a weaker pattern.

Burdens Rose Most Where High-Income Renters Grew Fastest

We took the top 50 US metros by total renter households[2] and analyzed the impact of each income group's growth on low- and middle-income cost burdens during both a pre-pandemic (2012 to 2019) and post-pandemic (2019 to 2024) window.

Since cost-burden filtering predicts that high-income demand matters most, we begin by grouping metros into two buckets: the top 25 by high-income renter growth (fast-growth) and the bottom 25 by high-income renter growth (slow-growth).

In both the pre- and post-pandemic windows, cost burdens for low- and middle-income renters rose more in fast-growth metros than in the slow-growth ones, producing a +2.3-percentage-point gap before the pandemic and a +2.5-percentage-point gap after it[3].

High-income renters show the same ordering but with smaller gaps. However, the cost burden gap between fast-growth and slow-growth metros narrowed for high-income renters after the pandemic (from +1.2 to +0.9 ppts) while widening for low- and middle-income renters (+2.3 to +2.5 ppts). The divergence further suggests cost pressures are filtering downward as the high-income renter population continues to grow.

High-Income Growth Stands Apart

Does high-income renter growth matter more to rising cost burdens than growth in other renter groups?

To find out, we ran the same test using the growth in middle-income and low-income renter households. We found that these variables produced much smaller or inconsistent differences between fast and slow metros than high-income growth.

Grouping metros by middle-income renter growth, the gap in cost burdens for low- and middle-income renters was -0.4 percentage points during the pre-pandemic window[4], an inversion of the gap observed when metros were grouped on high-income renter growth. Grouped by low-income growth, cost burden growth was flat in the fast-growth group but rose by 0.6 ppts in the slow-growth group during the pre-pandemic window, similarly inverting the high-income growth pattern.

During the post-pandemic window, grouping on middle-income renter growth did not produce an inversion of the fast and slow growth groups, but the gap between the two groups was significantly smaller than when grouped on high-income renter growth (+0.8 ppts compared to 2.5 ppts)[5]. Grouped on low-income renter growth during the same period, the cost burden growth gap was just +0.2 ppts.[6]

Grouping metros based on total renter growth had the closest similarities to our initial high-income growth test during the pre-pandemic window, producing +2.1 ppts[7] gap between fast-growth and slow-growth metros. However, we infer that this is because total renter growth includes high-income renter growth within its calculation.

The fast-growth versus slow-growth gap for metros grouped by total renter growth was smaller during the post-pandemic window_, at +1.6 ppts. Notably, testing these findings against a larger sample of the top 100 US metros by renter household growth produced similar results.

Cost Burdens Rose Most Where High-Income Demand Arrived

Timing was also an important factor to test. If today's high-income renter demand, not past demand, is driving today's cost burden increases, then burdens should shift when the metros leading high-income renter growth shift.

In both windows, metros that added more high-income renters tended to see larger cost-burden increases for low- and middle-income renters. However, the link was moderate.

During the pre-pandemic window, the correlation between high-income renter growth and low- and middle-income cost burdens was 0.56. The fast group added 10.4 high-income renters per 100 renter households, while low- and middle-income burdens rose. The slow group added 5.2 high-income renters per 100 while cost burdens for low- and middle-income renters fell.

The metros leading high-income renter growth shifted after the pandemic. San Francisco, New York, Chicago, Sacramento, Columbus, Kansas City and Oklahoma City each moved from the fast group to the slow group. In contrast, Phoenix, Tampa, San Antonio, Jacksonville, Baltimore, Tulsa and Fresno moved the other way. 36 of the 50 metros stayed in the same group during both windows.

In the post-pandemic window, the correlation between the two metrics fell to 0.41. The fast group added 4.4 high-income renter households per 100, and burdens for low- and middle-income renters rose. The slow group added just 0.1 high-income renters per 100, and burdens rose by less.

Crucially, the findings show that cost burdens for low- and middle-income renters followed high-income demand into new metros, but they did not leave when this demand slowed. Post-2019, cost burden increases accelerated by 7.6 points on average in the seven metros that joined the fast group, compared with 5.0 in the seven that dropped out and 5.1 in the metros that stayed put. Cost burdens rose where high-income renters arrived and kept rising where their growth faded.

Conclusion

Cost-burden filtering offers an alternative way to view the relationship between household growth and costs. The classical view of filtering typically describes a supply-side process in which, as homes age, they lose value and ownership of these assets filters down the income ladder. Stuart Rosenthal’s estimates in the American Economic Review using 1985 to 2011 panel data found a real annual filtering rate of 2.5% for rental housing.

Filtering also depends on demand. Rosenthal finds that filtering rates vary inversely with the income elasticity of demand and house price inflation. An analysis by Harvard's Joint Center for Housing Studies published in 2020 argues that when high-income renters absorb new construction, fewer older units filter down to lower-income renters. A subsequent 2026 review of America's rental housing further finds that between 2014 and 2024, the number of units renting for less than $1,400 fell by 9.3 million, while most new rental demand over the decade came from higher-income households.

Our analysis tests this demand channel directly. It asks: When high-income renters compete for rental units, do cost burdens rise in lower-income groups? While the findings don't prove causality, they show that in recent years, low- and middle-income cost burdens rose the most in places where high-income renter growth was the fastest, and much less so where low- and middle-income renter growth was.

Some important nuance and takeaways from the findings:

• This analysis presents an association across metros, measured in group averages, and it is not proof of causality.

• New York makes up roughly 26% of the slow group in the later window, a large weight in this group's results. Widening the test to the 100 largest metros cuts this to 23% while returning the same pattern in both windows.

• Our findings do not place blame on higher-income renters. Higher-income renters face limited supply too: JCHS found that from 2006 to 2016 high-income renter households grew by 2.9 million while units renting for $2,000 or more grew by 1.65 million. When supply falls short, renters across incomes compete for the same middle-priced units, and the pressure lands on those with the least room to absorb it.

• This analysis groups metros on high-income growth because cost-burden filtering predicts it. Sorting on middle- or low-income growth produced little of the gap, and sorting on total growth reproduced it only in part.

• The pressure follows high-income demand. Low- and middle-income cost burdens rose most in the metros where high-income renter growth was fastest, and burdens accelerated in the metros that gained high-income growth after the pandemic, though they kept rising where it faded.

  
  

[1] All income figures are expressed in 2010 dollars and are inflated across the analysis timeframe.

[2] The top 50 US metros remain the same in each period of the analysis.

[3] From 2012 to 2019, the share of low- and middle-income renters who were cost-burdened rose 1.2 percentage points in the fast group while falling 1.1 percentage points in the slow group. From 2019 to 2024, low- and middle-income burdens rose 5.9 percentage points in the fast group and 3.4 percentage points in the slow group.

[4] Here, cost burdens in fast-growth metros grew 0.2 ppts, compared with 0.6 ppts in the slow-growth group.

[5] Here, cost burdens in fast-growth metros grew 5.2 ppts, compared to 4.4 ppts in the slow-growth group.

[6] Here, cost burdens in the fast-growth metros grew 4.5 ppts, compared with 4.7 ppts in the slow-growth group.

[7] Here, cost burdens grew 1.6 ppts in the fast-growth group compared to 0.5 ppts in the slow-growth group.