Where Are All the Homes? Demographic Underpinnings of the Lack of For-Sale Inventory

by Dan McCue
Senior Research Associate

One of the challenges faced by housing markets has been a persistent lack of inventory of homes for-sale. Indeed, the most recent data on existing home sales from the National Association of Realtors® show that for 37 months we’ve been in a seller’s market – traditionally defined as a market in which there is less than six months’ supply of homes listed for sale (Figure 1). And, according to Redfin, this year’s spring buying season saw inventory nationwide hit record lows (see June Housing Markets Sets All-Time Records for High Speed and Low Supply).  In the many markets with few homes available for sale, new listings are almost immediately snatched up, with the high competition among buyers pushing prices out of reach of a growing number of would-be homeowners.    

Note: Data include existing single-family, condo, and co-op units for sale. Annual data are seasonally adjusted monthly averages. 2015 data are year-to-date through July. Source: National Association of Realtors®, Existing Home Sales via Moody’s Analytics.

There are several possible explanations commonly mentioned as to why for-sale inventories remain so tight, including the large number of owners stuck in homes because they are ‘underwater’ on their mortgages, the still-elevated volume of homes in the foreclosure process and held off the market, the lack of new construction in the years following the housing boom, and the many single-family units that have been taken out of the for-sale market to become rentals. But one additional reason not often discussed is demographics, which has also been playing a role in both the lack of inventory and in the slowness in new home sales over the past several years. Indeed, the ongoing generational shift among American households has slowed sales in the short run and is likely to continue to dampen sales over the next two decades.

Demographics and the Reduced Pool of Active Trade-up Homeowners

Over the past ten years, members of generation-X aged into the 30- and 40- year old age groups (Figure 2). As this relatively small generation, once called the baby-bust, replaced the large baby-boom generation now in their 50s and 60s, the population in their 30s and 40s declined. In some cases, the declines were stark. For instance, for the 35-39 year old age group, the population in 2013 was 9.3 percent smaller than it was 10 years earlier, with 10.4 percent fewer households.

Source: JCHS tabulations of US Census Bureau, Population Projections.

At the same time, as the 2015 State of the Nation’s Housing report mentions, the US homeownership rate took a significant dive, dropping to levels not seen in 20 years, with outsized declines among some age groups and the sharpest drop occurring among 35-44 year olds. Indeed, despite all the attention given millennials, homeownership rates among gen-Xers – particularly those currently age 35-44 – are actually furthest below 20-year historical rates of similarly aged adults (Figure 3).

Source: JCHS tabulations of US Census Bureau, Housing Vacancy Surveys.

So in combination, the demographically-driven decline in population of 30- and 40 year-olds was magnified by a sharp drop in homeownership rates, resulting in a significant decline in the number of homeowner households at these ages. Among the 35-39 year old age group, for instance, the number of homeowner households dropped fully 23 percent between 2003 and 13, and among 40-44 year-olds the decline was a substantial 19 percent (Figure 4). 

Source: JCHS tabulations of US Census Bureau, Current Population Surveys.

Traditionally, the 30s and 40s are key ages for housing market activity – particularly for trade-up and new home purchases. Indeed, homeowners aged 35-44 historically make up the majority of trade-up buyers (Figure 5). Fewer current homeowners in this key age group has meant fewer potential trade-up buyers and sellers, meaning fewer people putting their homes on the market, adding to tight inventories of for-sale homes. 

Source: JCHS Tabulations of American Housing Survey data.

Source: JCHS Tabulations of American Housing Survey data. 

Additionally, one of the most common ownership opportunities desired by trade-up buyers is a new home. Indeed, 35-44 year olds are also typically responsible for a high share of new home sales (Figure 6). And the majority of new homes sales to this age group are to those who are currently homeowners, so fewer current owners in this age group has also meant fewer potential buyers of new homes, fewer new home sales, and therefore a sizeable headwind to single family homebuilding. And there are other implications as well, such as for home improvements spending, given that most movers do some kind of post-move improvements, even if it’s just painting the walls, so fewer sales among gen-X has also affected remodeling spending markets as well.

Source: JCHS Tabulations of American Housing Survey data. 

While the millennials and baby boomers attract most of the headlines about how demographic trends are influencing housing demand, gen-X ers may actually be more influential than they get credit for in contributing to the recent weakness in single-family construction, home re-sales activity, and the widespread lack of inventory in many markets. 

One final note, however, is that the aging of the baby boom generation may also be contributing to the low levels of inventory and slower home sales – and this contribution may be a longer-lasting trend than that of the gen-X discussed above. Since mobility rates decline with age, the aging of the baby-boom will mean increasingly higher shares of older households who move less frequently. While there are concerns, most notably expressed in Dowell Myers’ insightful book Immigrantsand Boomers, regarding the potential future problem of elderly baby-boomers unloading a glut of housing on the market as they sell off or otherwise cease to head their own households, the oldest boomers are still only in their late 60s and so mostly many years from exiting the housing scene. And if for-sale inventories continue to remain tight as they are today, we may need to worry about the opposite problem: not enough turnover in the housing market to meet the needs of younger households, at least until boomers do reach the ages when they begin to vacate their homes in significant numbers. At present, it’s still hard to tell how much of the currently tight inventory is due to lingering effects of the housing downturn from longer term demographic shifts. Time will tell.  

Housing Cost Burdens Reach Higher Up the Income Scale, But Remain Nearly Unavoidable at Lower Incomes

by Ellen Marya
Research Assistant

In conjunction with the release of our 2015 State of the Nation’s Housing Report, the Joint Center mapped the prevalence of housing cost burdens – a key measure of housing affordability – in the US’s 900 metropolitan and micropolitan areas. As the map series illustrates, the share of households living in unaffordable housing varies dramatically across the country and for both owners and renters. Housing cost burdens, defined as the expenditure of more than 30 percent of household income on housing costs, are less pervasive in the country’s interior, while higher burden rates are largely concentrated in coastal and more urban areas and among renter households in particular. But within these nationwide patterns, high burden rates reflect unique local dynamics of household incomes and housing costs.

American Households Feel the Strain of Housing Cost Burdens 
(click map to launch)

The relationship between cost burdens, incomes, and housing costs within the 100 most populous metro areas is illustrated in Figure 1. As the figure shows, while housing costs and household incomes tend to rise together, the trend in cost burdens is somewhat less straightforward. Metro areas with lower cost burden rates (less than 30 percent of households with cost burdens, shown in yellow) are largely those with both low median housing costs and low to moderate median household incomes. Metros with moderate cost burden rates (between 30 and 35 percent of households with cost burdens, shown in orange) are more likely to have wider income distributions and slightly higher median housing costs, so that housing affordability becomes more difficult for those at the lower end of the income spectrum. This is even more striking in the group of metros with high cost burden rates (35 percent or more of households with cost burdens, shown in red), which includes several of the highest-income and highest cost metro areas; in these, affordability challenges move up the income scale.

Figure 1 (move cursor over figure to access additional information)

Notes: Housing cost burdens are defined as housing costs of more than 30% of household income. Households with zero or negative income are assumed to have burdens, while renters paying no cash rent are assumed to be without burdens. Source: JCHS tabulations of US Census Bureau, 2013 American Community Survey.

Indeed, cost burden rates among higher income groups rise dramatically as median housing costs rise (Figure 2). Among the large metro areas with the highest cost burden rates, the share of cost-burdened middle income households (with incomes between $30,000 and $45,000 annually) rises from just over one third in relatively low cost Tucson, to more than three quarters in high cost San Jose. A similar pattern holds among upper-middle and high income households, with cost burden rates topping 15 percent of high income households (earning more than $75,000 per year) in eight high cost metros.

Figure 2 (click on legend entry to display each income band)

Notes: Housing cost burdens are defined as housing costs of more than 30% of household income. Households with zero or negative income are assumed to have burdens, while renters paying no cash rent are assumed to be without burdens. High cost burden metro areas have a metro-wide cost burden rate of 35 percent or more. Metro areas are ordered from lowest to highest median monthly housing costs. Source: JCHS tabulations of US Census Bureau, 2013 American Community Survey. 

The persistence of housing cost burdens among higher income groups illustrates the lack of affordable housing options for even those with considerable means in some of the country’s most vibrant metro areas. Even more troubling, however, is the near ubiquity of housing cost burdens among lower income households. As Figure 2 shows, cost burden rates top 80 percent among households earning less than $15,000 per year – about equivalent to full-time work at the federal minimum wage – in all higher burden metros, even those with lower median housing costs. Additionally, the national maps show that no less than half of all households earning under $15,000 per year are housing cost burdened in every metro and micro area in the country, regardless of how low median housing costs fall.

The immense challenge faced by low-income households in finding affordable housing has been intensively detailed in a number of other analyses. In its annual Out of Reach report, the National Low Income Housing Coalition concludes that a full-time minimum wage worker cannot afford to rent a one- or two-bedroom apartment at Fair Market Rent in any state in the country, while the Urban Institute has mapped the growing shortage of units adequate, affordable, and available to lowest income renters in counties nationwide. As each of these inquiries shows, housing affordability remains a compelling need for the nation’s lowest income households.

How Much of the Damaged Housing Stock Was Rebuilt After Hurricanes Katrina and Rita?

by Jon Spader
Senior Research Associate

Ten years ago, Hurricanes Katrina and Rita created unprecedented damage in communities along the Gulf Coast. In addition to the human toll of the storms, the physical damage to the housing stock left many residents without a home to return to. On the 10^th anniversary, we now have a clearer picture of the extent to which homes damaged by the storms were eventually repaired or replaced with a new home.

In the months following the storms, FEMA conducted extensive damage assessments of residential properties to estimate the amount of damage that occurred during the storm. (While the FEMA assessment data are not exhaustive of every property that experienced hurricane damage following Hurricane Katrina, they are the most comprehensive source of information on damaged units.) In early 2010, a second assessment was conducted on a representative sample of these properties using a structured observation method, in which observers working from the street or sidewalk identified repair needs associated with hurricane damage, such as missing shingles and observable flood lines. Figure 1 shows the results of these observations for properties that experienced at least $5,200 in damage, the standard FEMA used to define “major” damage.

Notes: Estimates are representative of 1-4 unit residential properties that experienced more than $5,200 in damage from Hurricane Katrina or Rita. Source: Analysis of property observation data collected by Abt Associates.

These observations show that 17 percent of properties continued to show visible damage more than 4 years after the storm. A property was categorized as a ‘damaged structure’ if it showed one or more observable repair need and the observer did not deem the overall condition of the property to be good or excellent. Almost half of the properties still showing visible signs of hurricane damage (8 percent of all observed properties) contained structures that did not meet the Census definition for a ‘habitable’ structure. Under this definition, a housing unit need only be closed to the elements with an intact roof, windows, and doors, and no posted sign or other evidence that the property is to be condemned or demolished. The remaining properties contain a combination of rebuilt structures (70 percent) and cleared lots (13 percent).

Beyond these overall rebuilding rates, the data reveal clear differences in rebuilding outcomes across geographies and by tenure status. First, substantial variation exists in the percent of rebuilt properties across parishes, counties, and other subgeographies, ranging from 42 percent in MidCity Planning District to 96 percent in Jefferson Parish. These differences reflect a number of factors, including variation in the initial severity of damage and the resources available to residents to support rebuilding.

Within these geographies, properties occupied by homeowners had consistently higher rates of rebuilding than rental properties (Figure 2). Interpreting these differences is complicated by underlying differences in the siting, insurance coverage, and owner resources of homeowner and small rental properties. Nonetheless, some portion of the differences is likely attributable to the prioritization of homeowners in the programs established for providing rebuilding assistance. For example, in Louisiana, 59 percent of homeowner properties with major damage received Road Home rebuilding grants, compared to only 12 percent of rental properties. The average amount of the rebuilding grants provided to homeowners was $77,010. A more complete discussion of these programs and the allocation of rebuilding assistance is available in Turnham et.al. (2010).

The set of properties with damaged structures is also not evenly distributed across neighborhoods or properties. Instead, properties with remaining damage were frequently clustered together on blocks where at least one property contained a rebuilt structure. Sixty percent of owner-occupied properties with remaining damage—and 76 percent of rental properties with remaining damage—had a damaged structure on at least one of the two nearest properties on their block that also experienced hurricane damage. Yet very few damaged structures appeared on blocks that had been largely abandoned, containing only damaged structures or cleared lots. The resulting image is one of clustered pockets of remaining damage scattered among properties where other property owners returned to rebuild.

Taken together, these rebuilding outcomes highlight the extent of sustained damage more than four years after the storms. Today, it has been another five plus years since the property observations were conducted, so another round of observations might provide useful information about whether the damage remaining in 2010 was eventually resolved or whether it continues to appear on these structures today. In the interim, these estimates provide useful insight into the reconstruction of the housing stock following Hurricanes Katrina and Rita. More information is provided in Spader and Turnham (2014) and in an article in the forthcoming issue of Cityscape titled “Will My Neighbors Rebuild? Rebuilding Outcomes and Remaining Damage following Hurricanes Katrina and Rita.”

A City Revival? It Depends on Your Definitions

by Rachel Bogardus Drew
Post-Doctoral Fellow

Every spring, the Census Bureau publishes estimates of the population as of the prior July 1st at the sub-county level (i.e., individual municipalities, incorporated places, and non-incorporated county balances). Recent releases of these estimates suggest that cities have been expanding faster than their suburbs, with annual average population growth of 0.91 percent in 2010-2014 in the former, versus 0.77 percent average annual growth in the latter (Figure 1). Non-metropolitan areas, meanwhile, have recently seen their populations decline, by an average of more than 0.25 percent annually over the last four years.

Notes: Metropolitan areas and principal cities are defined as of 2013 by the Office of Management and Budget (http://www.census.gov/population/metro). Suburbs are all parts of a metro area not designated as a principal city. 
Source: JCHS tabulations of the U.S. Census Bureau’s Vintage 2014 Postcensal Population Estimates.

This recent trend of city populations growing faster than those of suburbs is a dramatic departure from prior decades, when suburban population growth significantly outpaced that of cities. Indeed, analyses by the Brookings Institution, using slightly different data and definitions of cities and suburbs but reaching the same general conclusions, suggest that between 2000 and 2010 the suburbs of large metropolitan areas grew by 1.38 percent per year, while the primary cities of large metro areas expanded by only 0.42 percent annually.

Some commentators herald this trend as a long-awaited sign of a grand urban revival, attributable to the combination of recent market events and long-running demographic trends. They posit that cities are becoming more popular in the wake of the late-2000s recession and downturn in housing markets, which exposed many of the downsides to suburban living. Either unable or unwilling to purchase homes in suburbs, the argument goes, more households are opting instead to live in urban neighborhoods with their mix of affordable housing options and lifestyle amenities. At the same time, some subsets of the population that are traditionally more likely to live in cities, including young adults, minorities, and childless households, have been increasing as a share of all households. Even if all households continued to live in cities and suburbs at the same rates as in the past, these demographic shifts alone would elevate the populations of cities more than those of suburbs.

Not everyone agrees with the urban revival hypothesis, however. Researchers point out that the suburbs are still home to about half of the U.S. population, and remain the location of choice for most households, including young families and retirees. Proponents of this position further argue that the cities that are growing the most do not resemble the dense cores that urbanists favor, but are instead places with lower densities and auto-centric commuting patterns more akin to close-in suburbs.

The disparity in these viewpoints exists in part because of the way that most analysts define cities and suburbs. Figure 1 above, for example, uses the federal Office of Management and Budget’s (OMB) classification of metropolitan areas and principal cities, and assumes that a principal city is effectively the ‘urban’ portion of the metro area, with the balance comprised only of ‘suburban’ places. This delineation, however, leaves no room for judgments about individual communities that blur the lines between urban and suburban areas. Some principal cities are themselves former suburbs that changed their status by making themselves more appealing to certain residents and/or by attracting a particularly large employer or industry, but that nonetheless retain many of their former suburban characteristics, such as neighborhoods with mostly single-family housing. Some suburbs, meanwhile, have developed dense centers with walkable amenities akin to cities, giving residents a taste of urban lifestyles in their small communities. The definitions used by OMB, however, group such places together with low-density bedroom communities under the broad umbrella category of suburbs. Indeed, even within a particular municipality there can be neighborhoods that have a more urban or suburban feel to them, so that classifying the entire municipality as either a city or suburb ignores the diversity that attracts residents to it in the first place.    

An alternative to the dichotomous definitions of city and suburb is to categorize places on one or more spectrums that take into account the different features of each type of community most significant to residents. These features can include everything from population density to type of housing stock, transportation usage, presence of retail establishments and cultural amenities, and walkability. Some researchers have already developed different classification schemes that take such characteristics into account, using density and commuting statistics to identify cities that are more or less urban in their character, or considering the degree of urbanization of suburban counties outside of large cities. Yet even these approaches still rely on assigned administrative boundaries of cities, counties, and metro areas that ignore the variations within them. A better option for those studying population shifts would thus be to see past political geography in favor of descriptive categories that better capture the diverse qualities sought by people as they choose where to live; researchers should define a spectrum of community types that represent this diversity. Such classifications will not only allow for richer analyses, but will better reflect the reality of where Americans live – not in cities or suburbs, but places that offer the best combinations of amenities to meet the needs of modern households.

Despite Declines in Homelessness, Family Homelessness Persists

by Irene Lew
Research Assistant

Since the 2010 release of Opening Doors, the first federal strategic plan to end and prevent homelessness, total homelessness in the US is now down 10 percent, from about 640,500 people in 2010 to 578,400 by 2014 according to HUD’s annual Point-in-Time (PIT) counts. This decline has been driven by a significant ramp-up in federal resources devoted to ending homelessness over the past decade. Although federal investment has resulted in substantial reductions in homelessness among at-risk groups such as veterans and individuals experiencing chronic homelessness, homelessness among families has persisted.  According to HUD’s most recent PIT count, more than a third (37 percent) of the total homeless population in the US is made up of people in families, with children under the age of 18 accounting for nearly 60 percent of this group. In fact, compared to other at-risk groups, homelessness among persons in families has declined at a much slower rate since 2007 (Figure 1).

Source: US Department of Housing and Urban Development, 2014 Annual Homeless Assessment Report to Congress: Part 1- Point-in-Time Estimates of Homelessness.

Most notably, much of the decline in family homelessness in recent years has occurred among the unsheltered population (those living on the streets, in abandoned buildings, vehicles or parks), while the number of sheltered homeless people in families (those living in emergency shelters, transitional housing programs, or safe havens) continues to rise steadily (Figure 2). Nearly nine in ten homeless people in families were staying in shelters in 2014. In New York City, where homelessness has reached historic proportions, the advocacy group Coalition for the Homeless estimates that homeless families make up the majority of homeless shelter residents and that the average number of homeless families in shelters rose by 67 percent between January 2005 and January 2015.

Source: US Department of Housing and Urban Development, 2014 Annual Homeless Assessment Report to Congress: Part 1- Point-in-Time Estimates of Homelessness.

Since the end of the recession, the affordable housing shortage has continued to play a major role in rising rates of family homelessness. Between 2010 and 2014, in high-cost locations where affordable rentals are in short supply, the number of homeless people in families increased substantially: by 50 percent in the District of Columbia, 41 percent in Massachusetts, and 22 percent in New York.  The problem is acute in urban areas across the country. According to the 2014 US Conference of Mayors Hunger and Homelessness Survey, 83 percent of the 25 cities surveyed cited the lack of affordable housing as a leading cause of homelessness among families with children in cities in 2014, with over a third (39 percent) of the 25 survey cities reporting that they expected the number of homeless families to increase moderately in the coming year. Indeed, family homelessness remains concentrated in urban areas, with 45 percent of all homeless people in families living in major cities in 2014. Nearly 20 percent of homeless people in families in the US lived in New York City, which had the largest concentration of homeless people in families in the country (41,633) in 2014, followed by Los Angeles City and County in a distant second (6,229). The concentrations of homeless people in families in New York City and Los Angeles also reflect the rising market rents in these metro areas, which have forced a growing share of households to allocate higher shares of their monthly incomes to housing costs. In 2013, 32 percent of renters in the Los Angeles metro area and 30 percent of renters in the New York metro area spent more than 50 percent of their monthly household income on housing, according to the American Community Survey.

The best  housing and services interventions for assisting homeless families have been a matter of some debate. Because families typically do not face long-term episodes of homelessness—indeed, HUD’s PIT count found that just 7 percent of homeless people in families were chronically homeless in 2014—one strategy often touted as suitable for helping them has been rapid re-housing, which focuses on quickly moving families out of shelters into permanent housing through the use of short-term rental subsidies. According to the National Alliance to End Homelessness, about three-quarters of families entering shelter are able to exit quickly with little or no assistance and never return.  However, results released from HUD’s Family Options study last month found that rapid re-housing was much less effective than a permanent housing subsidy, such as a housing choice voucher, in reducing shelter usage and improving housing stability of homeless families. Rapid re-housing may not work for all families, particularly those who are struggling with a host of long-term issues that may inhibit them from securing stable employment and achieving housing stability once their rental assistance expires.

For homeless families that require more intensive psychosocial support, permanent supportive housing, which pairs affordable housing with supportive services in order to achieve long-term housing stability, may be a more appropriate strategy. Through on-site services, permanent supportive housing can address the often complex causes of homelessness among families, such as histories of domestic violence, mental illness, and substance abuse. Yet the current inventory of permanent supportive housing largely targets single adults, especially those with chronic patterns of homelessness. Although the number of permanent supportive housing beds has increased significantly since 2007, a substantial share of permanent supportive housing beds are set aside for individuals rather than families (Figure 3).  The limited availability of subsidies for the services component, as well as higher operating expenses compared to affordable housing, present challenges for expanding the supply of permanent supportive housing. However, given the rising number of homeless families, it is important for policymakers, local communities, and practitioners to collaborate on interventions that address the continuum of needs that homeless families face, whether they be in the form of short-term rental subsidies, a housing voucher or permanent supportive housing. 

Source: US Department of Housing and Urban Development, 2014 Annual Homeless Assessment Report

to Congress: Part 1.