Kids in the neighborhood: The District has more children, but they are not where they used to be

After many years of decline, the total number of children in the District started to increase beginning 2011.


The turnaround in the overall population had happened earlier: the resident population bottomed out in 1998, remained flat for another eight years, and then it grew, and fast, especially since 2008, when we began adding over 10,000 net new residents each year. As we have shown here, most of the newcomers had been singles without children, who started replacing the families who had left the city. Only since 2011, the number of children began to grow, and in the last four years, the child population has grown at a faster rate than the adult population.


Separating children under the age of 6 and school-age children (ages 6-17) reveals differences in population trends.  The numbers for the younger children started recovering much earlier: the first positive numbers we see date back to 2005 and the growth has been steady since 2008 for this group. Among the older kids, however, the population slump began later than the younger kids (roughly 2001, compared to 1995 for children under 6), but it also recovered much later. So when we look at longer-term trends, we see that today we still have 3,500 fewer school-age children than we had in 1990, whereas we have 6,000 more children under the age of 6.


The familiar story we often hear—young people who moved into the city in early 2000s leave once their children hit school age—might be coming to an end.  Since 2011, we do not see—on net—the exodus of children from the city. But they now live in different neighborhoods.

To see the shift in where families live, we use the five-year summaries from the American Community Survey, which include data at the census tract level.  We use data from the 2010-14 summary, which combines five years of data and provides an average.  We then compare these to the 2006-10 data. Because the data are averages for five years, they don’t capture year-to-year variations.  For example, the five-year data summary does not show the recent growth in the number of school-age children, as it includes two years of decline (2010 and 2011), which overpower the three years of growth (2012, 2013 and 2014).  (But five-year-summaries are our only choice for this analysis. Because census tracts are small, the annual data are very unreliable; ACS does not even publish them.  Adding five years of data together makes things a bit more certain, but the error terms could still be large.)

We begin with where the children live. The map below shows the number of children in each neighborhood for the five-year period that ended in 2014. Most children are found in residential neighborhoods outside of downtown, especially in the neighborhoods just east of Rock Creek Park and at the southeastern edge of the city. The south east neighborhoods are among those that have the highest share of children in their populations: 30 percent in the Congress Heights, Bellevue, Washington Highlands cluster according to the 2010 census, and 34 percent in the Douglas and Shipley Terrace neighborhoods (though these numbers are well below their historical levels).  In contrast the share of children is under 20 percent in every neighborhood west of the park (the average for the city is 17 percent), except for Chevy Chase, where it is 23 percent (and increasing), 15 percent in Columbia Heights (where it is decreasing) and 20 percent in the Petworth cluster (where it has held somewhat steady). image016.png

Here is the interactive map.

How has this map changed since 2006? Below we show the change in the number of kids in each neighborhood in the last five years compared to the 2006-2010 period. You can see that that neighborhoods around the 16th Street and Georgia Avenue corridors north of Columbia Heights have attracted many families with children. The Takoma, Manor Park, and Brightwood Park cluster have added over 1,000 children, second only to the extremely popular Brightwood Park, Crestwood, Petworth neighborhood cluster, which added over 2,000 children. Capitol Hill now has fewer children (but only a modest decline of 100).   There are 1,280 fewer children east of the river now, but when broken down by neighborhood clusters, we see that five out of the 11 neighborhood clusters added more children (the biggest gains were in the Garfield Heights, Fort Stanton and Knox Hill cluster, which has seen lower vacancy rates and increasing home ownership).  Eastland Gardens and Kenilworth, on the other hand, lost half the children who lived there between 2010 and 2014 (a continuing trend for these neighborhoods), and the River Terrace, Benning, Greenway, Dupont Park cluster, which has been losing housing units and seeing increasing vacancy rates, also lost one third of the kids in the neighborhood.


Here is the interactive version of this map.

Families with young children make very different choices about where they live compared to families with school-age children.  They are more likely to move to the city, or stay in the city after having a baby, and could settle in any one of the neighborhoods. The map on the left below shows that one finds more children under the age of 6 in almost every neighborhood, but especially in the area that borders Rock Creek Park: neighborhoods along the 16th Street and Georgia Avenue corridors,  from Colombia Heights to Takoma Park, added  2,758 children under the age of 6. This is over a third of the net increase for this cohort. The notable exceptions to the growth are the neighborhoods that lie on the north west of the park  and a few neighborhoods east of the river, especially the River Terrace, Benning, Greenway and Dupont Park cluster. It is possible that young families are being priced out west of the park. Or perhaps, young parents are staying in the neighborhoods they lived in before they had kids.


Here is the interactive map.

Families with school-age children are pickier about where they settle. Neighborhoods north and west of downtown are adding more school-age children, neighborhoods that lie to the south and east of downtown continue bleeding.  There are more school-age children west of the river, but this is possibly aging-in-place as children who are included in the 0-5 group in the 2006-2010 data summary migrated to the school-age group in the 2010-2014 summary. Neighborhoods west of the park added 1,688 school-aged children. In comparison, the Brighwood Park, Crestwood, and Pethworth cluster, just by itself, added another 1,298. Columbia Heights, another strong destination for families with young children, cannot appear to hold on to the families with older children.  The cluster that holds Columbia Heights and Mt. Pleasant lost 574 children between the ages of 6 and 17. This is also true for the Brookland, Brantwood, and Langton cluster.

The biggest declines are in the neighborhoods east of the river, which collectively lost 2,654 school-age children. Losses were greatest in the neighborhoods along the northern border with Maryland.  The neighborhoods in these clusters, similar to the south east neighborhoods, used to have very high concentration of children but have been losing them since 1980s. For example 35 percent of the Eastland Gardens and Kenilworth’s population are children, down from 39 percent in 2000. In the Deanwood, Burrville, Grant Park, Lincoln Heights, and Fairmont Heights cluster, the number of children in 2010 was half of what it was in 1980. This neighborhood is now adding younger children, but not yet, school-age children. And given housing prices in the city and household incomes east of the river, most families who left these neighborhoods have likely moved out, and did not relocate in another neighborhood in DC.

The population dynamics across neighborhoods provide yet another picture of gentrification and where we are most likely to find it.  The city’s most expensive neighborhoods (when it comes to housing) are holding on to the school-age children but are not able to add young families. While east of the river continues to have the highest concentration of children, if trends continue, neighborhoods near the 16th Street and Georgia Avenue corridors could claim this distinction soon. Neighborhoods east of the river are adding younger children, but rapidly losing school-age children, and on the net losing families.

What exactly is this data?

There are 39 neighborhood clusters in the District, and unlike wards or census tracts, they are not drawn to be of similar size. Some are very small, others are large and densely populated. For example, the Brightwood Park, Crestwood, Petworth neighborhood cluster has the greatest number of children (8452), but it also has many adults, so the share of children in its population is only 20 percent.

We compile the neighborhood data from the five-year ACS data summaries, which include census-tract level data, using the tract-to-neighborhood cluster mapping from

Data on child population is from


How can the rent be so high in DC when almost two-thirds of all rental units in the District are subject to rent control? A small number of “spoiler “units with high turnover may be the reason.

The District, along with a few other housing major markets in the nation, has rent control laws that were enacted to protect tenants against unreasonable rent increases. The laws governing rent control in these markets generally stipulate that rent increases are bound by the Consumer Price Index or other cost of living measures. These laws also allow for larger increases when units become vacant (we will return to this point later).

If rent increases are generally constrained by the cost of living, how could the median rent in the District over the 2005-2014 period have grown by 65 percent, or nearly 6 percent annually, when the DC consumer price index over the same period grew by only 30 percent, or just under 3 percent per year? In addition, as shown below, growth in median rents in DC has outpaced other markets where rent control laws are also in place.

Median Contract Rent: 2005-2014


Source: U.S Census Bureau, American Community Survey, DISTRICTMEASURED.COM

One possibility is that most units are not subject to rent control and must pay what the market will bear. To analyze this, we looked at data on the number of rental units subject to rent control in the District. Under DC laws most rental units with more than 5 units and built prior to 1975 are subject to rent control.

Here’s what we found in terms of the number of units subject to rent control



Source DC Office of Tax and Revenue, DC ORA, DISTRICTMEASURED.COM

This is what the data looks like by Ward


Source: DC Office of Tax and Revenue, DC ORA, DISTRICTMEASURED.COM


  • Almost two thirds of all rental units are potentially subject to rent control or other restrictions. This is a significant share of the entire rental stock.
  • The highest concentration of rent control units was located in Wards 3 and 4 with over 80 percent of all units potentially subject to rent control.
  • Wards 5, 6 and 7 had the lowest overall share of rent controlled units. In Ward 6, less than a third of all units are subject to rent control.
  • These overall findings are largely consistent with the results of a prior study from the Urban Institute, “A Rent Control Report for the District of Columbia” by Peter A. Tatian, Ashley Williams, June 17, 2011 which can be accessed by clicking here


To put these statistics further into perspective we looked at the share of rent regulated apartments in New York City.

This is what we found for New York City’s Boroughs


 Source: 2014 New York City Housing Vacancy Survey, U.S. Census Bureau

DC ‘s overall share of rent regulated units is comparable to New York City’s, although certain DC Wards have an even higher share of regulated units than in the Bronx or Manhattan.

So what accounts for the large increases in rents given that two thirds of DC rental units are subject to rent control?

Two provisions of the law are likely to account for this:

When a DC tenant vacates a rental unit the amount of rent charged may, at the election of the housing provider, be increased:

(1) By 10% of the current allowable amount of rent charged for the vacant unit; or

(2) To the amount of rent charged for a substantially identical rental unit in the same housing accommodation; provided that the increase shall not exceed 30% of the current rent charged for the vacant unit.

It is easy to see how the combination of these two provisions can result in substantial price increases. This is what could happen to rents in a building where one unit (Unit 1) has a relatively high turnover. We assumed a turnover of once every three years for this unit, which is not unusual given DC’s high mobility. The other units have no or limited turnover. We assumed all comparable units started with a rent of $1,000 in 2004.

Rent simulation given turnover


Source: DC Office of Revenue Analysis, Cells highlighted in orange indicate a vacancy rent increase of 10%, and red denotes up to a 30% percent increase based on a comparable unit

Turnover in one “spoiler” unit can cause rents to increase for all comparable units in the building.

Even the third unit in this hypothetical building, which turns over only once in ten years, has seen its potential rent increase by 58 percent. Only the fourth unit, that has seen no turnover, has a rent that remains below $1500.

As shown above, given allowable vacancy increases and comparability under DC law, even one comparable unit with a high turnover can cause rents to increase substantially for many units. Higher turnover, which may be due to changing demographics (more married couples and fewer singles remaining in the same unit for many years) or a spoiler unit, which may be the substantially identical unit on the same floor but close to the garbage room or near the garage exit, could cause rents to increase significantly even with rent control.

What exactly is the data?

To determine units potentially subject to rent control we looked at the year built and number of units for the following building codes for rental properties (21, 22, 25, 28,216 and 217) from the DC Office of Tax and Revenue Real Property Tax Database. The New York City Housing and Vacancy Survey (NYCHVS), sponsored by the New York City Department of Housing Preservation and Development, is conducted every 3 years to comply with New York state and New York City’s rent regulation laws. The Census Bureau has conducted the survey for the City since 1965. The 2014 NYCHVS is the 16th such survey. No similar study exists for DC.

Bob Zuraski contributed to this post

State and Local Household Tax Burdens: DC’s are the Lowest in the Metro Area

In hypothetical comparisons, tax burdens on the District’s families are lower than those in surrounding jurisdictions in the Greater Washington Metropolitan Area. This is according to a recent study we released, Tax Rates and Tax Burdens 2014: Washington Metropolitan Area, that compares tax burdens across four main taxes: income, property, sales, and auto.  In this study, we take a dual-income family with one child and calculate the taxes it would have to pay across the metro area. We compare each family’s DC tax burden with the tax burden it would face in the surrounding jurisdictions, including Montgomery and Prince George’s Counties in Maryland, and Arlington County, Alexandria, Fairfax County, Fairfax City, and Falls Church, in Virginia. We repeat this calculation at five different annual income levels: $25,000, $50,000, $75,000, $100,000, and $150,000.

Click on the graphic below to interact with the map and filter by tax type and income level.

2014 Metro Map SnapShot

These hypothetical comparisons show that DC’s combined tax burden ranks lowest at each income level. For the family earning $25,000/year, DC’s combined burden for all four taxes was $3,288, or 13.2 percent of income, while the combined DC tax burden was $12,747 for the family earning $150,000/year, representing 8.5 percent of its income. The regional average combined tax burden is 15.1 percent for the lowest income family, and 9.6 percent for the highest.

Assessing how the other areas rank depends on the specific income level. At the lowest income level of $25,000/year, the Virginia jurisdictions levy the highest combined tax burdens, with Alexandria, VA, ranking first overall, and levying a tax burden of $4,130, or 16.5 percent of the family’s income. At the highest income level, Falls Church, Virginia, levies the highest overall tax burden, at $15,625 or 10.4 percent of income. (See the bar charts at the end of this post to view the composition of each jurisdiction’s combined tax burden by tax type and income level.)

Income Tax

The District’s income tax burden is lower than the metropolitan average at all levels except at the $100,000 and $150,000 income levels. DC’s income tax burden was a negative $732 for the family earning $25,000, meaning that the family would receive a refund through the Earned Income Tax Credit. The family in Montgomery and Prince George’s Counties at this income level would receive a slightly smaller refund, and the family in any of the Virginia locations would owe $0 in income tax payments.

At each of the other income levels, the income tax burdens in the District fall in the between the Virginia and Maryland jurisdictions. The income tax burdens in Virginia are the lowest, and the income tax burdens in the two Maryland Counties are the highest because they each levy an additional county-level income tax of 3.2 percent.

Metro Area Income Tax Burdens, in $, All Income Levels

Source: ORA Analysis.

Property Tax

Real property tax burdens for District of Columbia residents fall below the area-wide averages at all income levels, except at the $25,000 income level. We assume that the family at this income level rents an apartment, and that the property tax portion of rent comprises about 20 percent of rental payments.

The data we use for calculating property tax burdens, both median rents data from the US Department of Housing and Urban Development, as well as median home value and median income data from the US Census, are aggregated at the level of the Metropolitan Statistical Area (MSA). Since all of the metro area jurisdictions are in the same MSA, the house values used for calculating the property tax burdens are the same across the region. This means that the calculated property tax burden for renters is the same for each jurisdiction, while the property tax burdens for the other four income levels reflect the differences in real property tax rates and property tax relief provisions in each locale, and not the housing sub-markets in each area.

In 2014, the median income of mortgage-holders in the Washington, DC, MSA was $122,129 and the median house value was $389,700. Using a multiplier based on these values, we calculated a scaled house value for each income level in our study.

See the graphic below for the property tax burdens, by income level. As the chart illustrates, the property tax burdens rise linearly with income (with the built-in assumption that house values also rise with income), for the families assumed to own their homes (those with incomes above $50,000/year). There is an exception for the family earning $50,000 in Montgomery County, MD, because it is eligible for a supplemental homeowners credit for lower income homeowners.

Metro Area Property Tax Burdens, in $, All Income Levels

Source: ORA Analysis. Note: calculation for family earning $25,000 (assumed to be renters) is different than the calculation for the other four income levels (assumed to be homeowners).

Sales Tax

The District levies a general 5.75 percent sales tax, the lowest of the jurisdictions in the study, but has a higher calculated sales tax burden than the metropolitan area average at all income levels. This is due to higher sales tax rates on consumption goods such as restaurant meals, parking, hotels, and rental cars. These are items that are more likely to be purchased by tourists, or non-resident commuters, but the higher rates apply to District residents, as well. We are not able to separate the consumer spending of tourists and commuters from the expenditures made by a jurisdiction’s residents.

Virginia has a 5 percent general sales tax and each of the jurisdictions in the study levy an additional one percent local sales tax, for a total 6 percent rate. Sales tax burdens at each income level are the same in Alexandria, Arlington, Fairfax City, and Falls Church, and represent the second highest burdens behind the District, partly due to higher levies on hotels and restaurant meals. Maryland has a 6 percent general sales tax, which applies in Montgomery and Prince George’s Counties. Both Maryland counties have the same sales tax burdens at each income level and are the lowest in this study. The Maryland jurisdictions do not have a higher restaurant meals tax, like the District and several of the Virginia jurisdictions.

Auto Tax

Virginia localities have the highest auto tax burdens at all income levels, with Alexandria’s ranking highest. At the $25,000 and $50,000 income levels, Montgomery and Prince George’s Counties in Maryland had the lowest auto tax burden, only $5.00 lower than the District. At the $75,000 income level, the District’s auto tax burden also was slightly higher than the Maryland Counties’ burden of $374. The District and Maryland jurisdictions do not levy a personal property tax on automobiles, as in the Virginia localities.

See the full report for more results and a description of the methodology.

The follow charts provide comparisons of the tax burdens from each tax, by jurisdiction, at each income level.


Income = $25,000

Source: ORA Analysis.


Income = $50,000

Source: ORA Analysis.


Income = $75,000

Source: ORA Analysis.


Income = $100,000

Source: ORA Analysis.


Income = $150,000

Source: ORA Analysis.


What exactly is this data?

Because the District levies taxes typically employed by both state and local governments, the study compares DC’s tax burdens to the combined state and local tax burdens the family would face if it lived in the neighboring jurisdictions. This is the same methodology used in our annual Tax Rates and Tax Burdens 2014: A Nationwide Comparison, which compares DC tax burdens for a hypothetical family to those that would apply if the family lived in the largest city in each state in the US. [See our recent post highlighting those results here.]

We calculated the tax burdens using housing value and median income data from the US Census Bureau’s American Community Survey; median rent data from the US Department of Housing and Urban Development; income profile data from the Internal Revenue Service’s DC Statistics of Income; consumption information from the Bureau of Labor Statistics’ Consumer Expenditure Survey, and auto data from the National Automobile Dealers Association Used Car Guide, the Environmental Protection Agency, and the Department of Energy. Tax rate data were obtained from state and local tax officials, state and local government web sites, and third party sources such as the Federation of Tax Administrators and CCH/Wolters Kluwer.