|
School-Age Children in Multifamily Homes? Very Few.
If you build apartments, it leads to school overcrowding, right? That’s one of those long-standing notions that sounds reasonable – until you look at the numbers. And this month, there’s a new batch of numbers to look at.
Every two years, HUD and the Census Bureau release a new version of the AHS, which tracks not only housing characteristics, but other related questions as well. NAHB Multifamily has used those survey results extensively, mining them for data about children in various types of housing units. Results published previously in NAHB Multifamily Market Outlook (November 2001 issue) showed that for every 100 single-family households, there are about 62 school-age children, while the average number is much lower — about 37 — for multifamily households.
The Census Bureau has just released a new set of data from the most recent (2003) AHS. The updated data confirm the 2001 findings. Most striking: On a per household basis, multifamily has fewer children than other types of housing structures, and the number of school-age children is particularly low in very large multifamily structures, in multifamily condos, and in new multifamily developments.
This article also addresses some new questions that have arisen since the 2001 article was published. For example, the new data show that, in apartment buildings with at least five units, the number of children present is strongly related to the number of bedrooms. And among the two-bedroom apartments in those buildings, the number of school-aged children is somewhat lower in the more upscale units.
Each New Student Costs Money
The cost of public schools generally is the largest expenditure in local government budgets. According to the Census Bureau’s Governments Division, local governments in the U.S. spent a little over $1 trillion in total during fiscal year 2000-2001, and $411 billion of that was spent on education (the lion’s share, $388 billion, going to elementary and secondary schools). In contrast, they spent less than $25 billion on an item such as housing and community development (Figure 1).
Although residential development has an impact on the demand for public schooling, it is only one of several factors. Even if no home building takes place, school enrollments may rise as a result of natural population increase among nonmoving households, or as households with different numbers of school-age children move into and out of existing homes, apartments and condos.
In most parts of the country, other factors further tend to mitigate the impact of new construction on local school systems. For one, when new households move into an area, school districts typically receive additional state and federal government aid.
For another, some households send their children to private schools. According to the National Center for Education Statistics (NCES), the share of students educated in private schools is 12.6% of all school-aged children nationally. Since public money rarely is used to pay for private instruction, this tends to further reduce public school expenses, although the extent to which that occurs varies from place to place. The NCES also shows that 1.7% of students aged 5 to 17 receive home schooling, which further reduces the cost of public education per household.
Also, when the issue is the property tax burden borne by a typical homeowner in a community, the average cost per pupil becomes relevant. If there is excess capacity in the school system or economies of scale otherwise exist, cost per pupil actually may go down as enrollment rises. A 1995 study by Fabio and Sonstiele found evidence that this sometimes occurs. 1
Does it Matter Where the Children Live?
The 2003 AHS shows that, for the U.S. as a whole, there are 54.2 school-age children (ages five through 18) for every 100 households (Table 1).
But this varies substantially depending on structure type. In multifamily units, there are only 36.9 school-age children per 100 households — compared to 62.4 for single-family detached, 47.0 for single-family attached, and 53.8 for manufactured housing (Figure 2). A traditional explanation for the differences is that families with school-age children have a particularly strong demand for homes with their own back yards.
Source: 2003 American Housing Survey, U.S. Census Bureau and the Department of HUD
The number of school-age children tends to be even lower in larger apartment buildings — only 26.1 school-age children per hundred units in structures with at least 20 apartments, compared to 47.4 in 2-4 unit structures.
There also are some interesting relationships between structure type and other characteristics. In particular, within a given structure type
- renters have more children than owners,
- recent movers have more children than households who stay put, and
- households moving into existing homes have more children than those moving into brand new units
But the magnitude of the differences varies according to the type of structure. In single-family detached, for instance, the mover vs. non-mover effect is very strong compared to the new vs. existing difference. The result is that new single-family detached homes tend to attract households with more children than current residents of single family detached have on average — 73.8 per hundred vs. 53.4.
In multifamily, on the other hand, the new vs. existing unit effect is much stronger, so that households moving into new multifamily apartments have fewer children than non-moving apartment dwellers — 26.5 per hundred vs. 33.9. The implication is that a local school district may overestimate how many school-age children are going to be in a new multifamily structure during the first year of its life if it bases the estimate on the number of children per household in existing structures.
Also, there tend to be fewer children per household in larger apartment buildings. Households in 2-4 unit buildings have 47.4 children per hundred, on average, while households in 20+ unit structures have only 26.1. If this is combined with the effect of ownership status, it produces a very large spread between nearly 50 children per 100 renters in 2-4 unit multifamily buildings and fewer than 12 per 100 owners of condos in 20+ unit buildings (Figure 3).
Source: 2003 American Housing Survey, U.S. Census Bureau and the Department of HUD
If the effects of number of units in the building, age of the structure, ownership status, and recently-moving households all are taken into account, the highest number of children per 100 households is 89.2 among 100 households that have recently moved into single family detached rental units. This suggests that households with school-age children are more likely to trade off a chance to own or live in a newer structure in order to get the back yard.
The lowest number is 10.1 among households who bought condos recently in 20+ unit buildings.
Number of Bedrooms is Important
For multifamily housing, separate local development standards sometimes exist for units with different numbers of bedrooms. This is possible largely because most apartments fall into one of only a few bedroom categories. According to the AHS, of the occupied rental apartments in five-plus unit buildings, almost 90% have either one or two bedrooms.
Not surprisingly, there is a strong correlation between number of bedrooms and number of school-age children in the unit (Table 2). Five-plus apartments with fewer than 2 bedrooms contain, on average, only 10.4 children per 100 households, while apartments with at least three bedrooms in similar buildings have 93.4 — even more than households in single family detached. The small share of apartments with more than two bedrooms prevents this from pushing the overall average for multifamily up higher than it is, however.
At the other end of the scale, zero-bedroom apartments also make up a small share of the market, making separate tabulations for them difficult with a survey the size of the AHS. To extent that the tabulations are possible, they show little difference between the numbers of school-aged children in efficiencies and one-bedroom apartments.
Many of the effects of ownership, recent-mover status, and age of the structure on number of children persist after controlling for bedroom size. The differences between households that have moved recently into new and existing units, for example, remain apparent (Figure 4).
* The figure for recent movers into new 5+ multifamily units with at least 3 bedrooms is based on only 16 observations. For that reason, it should be interpreted with caution.
Source: 2003 American Housing Survey, U.S. Census Bureau and the Department of HUD
Fewer Children in Upscale Multifamily?
On several occasions, NAHB has heard speculation that apartments at the high end of the market tend to contain fewer children. The AHS provides some evidence to support this, although only in certain cases.
There is no universally accepted way to define quality submarkets for rental apartments. The approach adopted here is to differentiate five-plus apartments into different categories based on rent. Gross rent (which includes all utility payments except for telephone service) is used, because it is less ambiguous than contract rent, due to differences in the way utility payments are handled in different projects.
To account for the fact that a rent of $1,000 per month means something different in downtown Manhattan and rural Mississippi, gross rent is scaled to Area Median Family Income (AMI). Four market quality segments — low, medium-low, medium-high, and high — are defined based on gross income as a fraction of AMI. The cut-offs are 13.2%, 16.6%, and 20.4% of AMI, respectively. In a location with an AMI of $60,000 (which works out to an even $5,000 per month and is fairly close to the $57,500 AMI for the U.S. as a whole in 2004), the cut-offs fall at monthly gross rents of $660, $830, and $1,020.
These cut-offs are based on quartiles — that is they are chosen to include approximately equal numbers of 2-bedroom units in each rent tier. It’s necessary to control for bedroom size, since the positive correlation between rent and number of bedrooms obscures the results otherwise. The AHS sample contains too few observations to permit a separate analysis of anything other than one- or two-bedroom apartments.
Among the one-bedroom apartments, there appears to be no systematic relationship between school-aged children and rent. The two-bedroom apartments show a relatively clear pattern, however. Starting at the low end, there is a slight increase in number of school-aged children as we move from low-rent to medium-low rent apartments. After that, the number of children falls off somewhat, so that there are only 37.6 children per hundred households in the high rent category, compared to 51.3 in the bottom tier (Table 3).
Anyone interested in applying the results in Table 3 to a particular area can easily find the current AMI for that area on the HUDUSER web site, at http://www.huduser.org/datasets/il/il04/index.html.
Good Housing Info Leads to Better School Census Estimates
Although new construction has an impact on demand for local schools, there are other factors at work as well — such as households with different characteristics moving into existing units. In most cases, households moving into new housing units have fewer children than households moving into existing units. This effect is particularly strong among renters, and among occupants of multifamily in general.
In the U.S., on average, there is less than one school-age child per housing unit, but there are fewer children in multifamily than in other types of housing, and fewer still in particular types of multifamily units such as condos, or units that have fewer than two bedrooms, or are in buildings with more than 20 units.
Education is a large item in the budget of local governments. Those governments need to take the increased costs of providing education into account when planning for new residential development — but they also need to do so as accurately as possible. Failure to take many of the factors described above into account can lead to overestimates of the costs, particularly for multifamily development.
1 Fabio and Sonstiele, “Did Serrano Cause a Decline in School Spending?” National Tax Journal, 1995.
[
return to top ]
|
