Introduction

The 1990 Decennial Census of Population and Housing is the single most important source of data on the physical and social characteristics of housing in Virginia. These data allow us to update our understanding of the current status of housing within the Commonwealth and to assess trends established since the 1980 and earlier censuses. The Housing Atlas for Virginia has two primary purposes: to provide an easily accessible description of housing in Virginia by words, charts, and maps, and to provide a statistical profile of housing in 1990 for the state, Planning District Commissions (PDCs), and each county and independent city. The boundaries of PDCs are given in Figure 1.

The Atlas is organized into four major sections: Text, which reviews long-term trends in housing and the geographic patterns of housing across the state; Maps, providing a graphic presentation of these housing patterns; Profiles, a one-page compendium of the most important 1990 housing data released by the Census; and Data, a listing of the values for each of the variables graphically presented in the Maps section.

The Text and Profiles are further organized into four sections: Housing Unit Characteristics; Financial Characteristics; Housing Quality; and, Household Characteristics. Obviously this Atlas can only present a small portion of the data available from the 1990 and earlier censuses, and there are sources of housing data other than the Decennial Census. In addition, this Atlas is not intended to provide a detailed assessment of housing needs within Virginia, although the data presented certainly document the magnitude and location of many of the state's housing problems. Such assessments are available in official housing plans (known as Comprehensive Housing Affordability Strategies, or CHASes) prepared by the state and many local communities.

Figure 1: Virginia Planning District Commissions is available from the server.The Virginia Center for Housing Research (703-231-3993) is available to assist you in identifying and accessing housing data from the Census and other sources. In addition to the Virginia Center for Housing Research, housing (and other) data from the 1990 Census are available from the Virginia State Data Center in Richmond (804-786-8624), the Center for Public Service at the University of Virginia (804-982-5522), and the Planning District Commissions.

Readers should note that the universe (the group covered by any given data item) for all tables and figures is provided with the table or figure. The universe for each map is provided in a detailed discussion of sources of data at the end of this section. If a discrepancy is found between the data provided in the Atlas and Census data obtained from other sources, it could be due to differences in the universe covered. Any questions or comments about the Housing Atlas should be directed to the Virginia Center for Housing Research.

Housing Unit Characteristics

There were 2.5 million housing units in Virginia in 1990, an increase of 475,393 units (23.5 percent) over 1980. The state's housing stock has grown substantially over the past five decades, particularly after 1960. As shown in Figure 2, single-family detached units have not only been the most popular housing form, their popularity has increased with each passing decade. Single-family attached units (or townhouses) have also increased substantially over the past two decades, as have mobile homes. By 1990, single-family units (detached, attached, and mobile home units) made up 77 percent of the state's housing stock. These single-family units provide the bulk of the state's owner-occupied stock, but are also an important component of the rental inventory.

The supply of multi-family housing has increased much less rapidly than single-family detached units. The supply of housing units in duplexes has decreased from a peak of 83,378 units in 1950 to only 53,310 units in 1990. This loss has been offset somewhat by an increase in units in structures with 3 or 4 units from 48,355 in 1950 to 73,709 in 1990. However, the growth in the multi-family sector has been almost entirely in structures with 5 or more units, which have had a growth rate since 1960 roughly on par with town-houses and mobile homes but less than that for single-family detached units.

The range in growth rates in housing markets throughout Virginia from 1980 to 1990 was enormous: from a loss of 12.5 percent to a gain of 85.7 percent. The housing supply increased by 25 percent or more in several major growth regions (see Map 10): Northern Virginia, the Charlottesville-Richmond metropolitan areas, and the Norfolk metropolitan area. In addition, Frederick and Shenandoah counties in the northwest corner of the state and Bedford and Franklin counties toward the southwest were smaller growth centers.

Although there was an increased concentration of the state's housing in metropolitan markets (the boundaries of which are shown in Figure 3), within these markets the pattern has been toward dispersion away from urban centers. With a few exceptions, the supply of housing in central cities throughout the

Figure 3: Virginia Metropolitan Statistical Areas is avaiable from the server.state declined or was virtually stagnant while the supply of housing in the suburban and exurban (rural areas on the fringe of metropolitan housing markets) locations increased rapidly.

Many of the rural areas of the state, particularly in the southwest and along the North Carolina border, experienced declines or very slow growth in the supply of housing. Nonetheless, 1990 owner and renter vacancy rates for Planning District Commissions in these areas were generally lower than the state average. (See Table 1.)

State and PDCs Vacancy Rate Vacancy Rate

(Owners) (Renters)

Virginia		2.1	8.1
Lenowisco		1.3	8.7
Cumberland Plateau		1.4	8.2
Mount Rogers		1.4	6.6
New River Valley		1.6	5.9
Fifth		1.6	7.8
Central Shenandoah		1.4	4.4
Lord Fairfax		2.2	5.8
Northern Virginia		2.2	8.6
Rappahannock-Rapidan		2.4	4.5
Thomas Jefferson		1.5	4.4
Central Virginia		1.6	7.5
West Piedmont		1.2	7.4
Southside		1.3	5.4
Piedmont		1.1	5.0
Richmond Regional		2.0	8.8
RADCO		2.5	6.9
Northern Neck		2.1	7.5
Middle Peninsula		2.1	7.5
Crater		1.5	6.9
Accomack-Northampton		3.0	9.1
Hampton Roads		3.3	9.2

Vacant Units (for rent or for sale) divided by vacant units (for rent or sale) + occupied units (renter or owner).

Perhaps no form of housing supply demonstrates as much geographic variation as mobile homes, which are scarce in cities and in the suburbs of larger, metropolitan areas, but abundant in the southwest and southern portions of the state. (See Map 14.) Although these same areas had significant increases in the supply of mobile homes between 1980 and 1990, there were several clusters of counties where the number of mobile homes increased two-fold or more. (See Map 15.) These include Fauquier and Spotsylvania counties in the north, several counties in the Northern Neck and Middle Peninsula Planning District Commissions along the Chesapeake Bay, a cluster of counties in the Hampton Roads and Crater PDCs, a group of counties stretching from Augusta to Mecklenburg, and a group in the Bedford-Martinsville areas.

The highest concentrations of multi-family housing are in the Northern Virginia, Richmond Regional, New River Valley, Fifth (Roanoke Valley), and Hampton Roads PDCs. (See Table 2.) More than one-in-five housing units in these areas are in multi-family structures.

Due to significant differences in the periods during which the housing supply was built throughout the state, some areas are distinguished by a preponderance of new housing and others by older units. For the most part, the newest housing is found in the counties that are the suburbs of fast growing metropolitan areas and the oldest housing is found in central cities. This is shown in Map 11, which gives the median year residential structures were built in each community; Map 12, which gives the percent of units built between 1980 and March, 1990; and, Map 13, which gives the percent of units built before 1940. Except for a few rural counties, central cities are the areas where housing had a median age of 25 years or older in 1990 (i.e., where the median year residential structures were built was before 1965). In contrast, several suburban areas had

State and PDCs Single-family 2-4 Units 5 or More Mobile Home

Units Units & Other

Virginia	1,748,056	143,530	422,648	182,100
Lenowisco	26,536	1,134	1,382	8,986
Cumberland Plateau	32,956	1,027	1,445	14,365
Mount Rogers	56,373	3,075	3,910	13,177
New River Valley	38,288	3,388	10,127	8,806
Fifth	77,563	9,316	16,523	4,939
Central Shenandoah	65,954	6,617	7,576	9,609
Lord Fairfax	53,278	3,288	5,041	5,927
Northern Virginia	389,676	13,482	172,826	7,601
Rappahannock-Rapidan	38,389	1,550	1,867	2,930
Thomas Jefferson	47,549	4,657	9,414	6,455
Central Virginia	59,704	5,722	6,937	11,655
West Piedmont	72,824	4,943	6,926	17,595
Southside	26,095	849	851	8,037
Piedmont	25,041	824	634	6,942
Richmond Regional	216,905	22,118	57,971	9,137
RADCO	48,850	1,639	4,609	6,549
Northern Neck	20,288	373	286	3,369
Middle Peninsula	26,341	932	848	5,505
Crater	44,912	5,653	5,248	6,068
Accomack-Northampton	16,717	617	359	4,330
Hampton Roads	363,817	52,326	107,868	20,118

Universe: Total Housing Units

residential structures with a median age of 15 or fewer years: many of the suburban counties in Northern Virginia; Frederick County (outside Winchester); Powhatan and Chesterfield counties (outside Richmond); Gloucester, New Kent, and James City counties (outside Williamsburg); Virginia Beach and Chesapeake cities; and Bedford County.

With the exception of central cities and some counties near the mouth of the Rappahannock River (other than Lancaster County), 25 percent or more of the housing units in most of the Fairfax-Charlottesville-Richmond-Norfolk area were built between 1980 and March, 1990. Newer units are also prevalent in Frederick, Shenandoah, Warren, Nelson, Campbell, Bedford, Franklin, and Montgomery counties. Except for most cities, the communities with the highest percentages of pre-1940 housing are in the Central Shenandoah, Lord Fairfax, Rappahannock-Rapidan, and Accomack-Northampton PDCs. This includes several exurban communities that have relatively high percentages of both pre-1940 and post-1979 housing stock.

Financial Characteristics

Being a homeowner is often identified as a vital part of the American Dream. Homeownership is largely a function of income, life-cycle, and cost. Few young adults are homeowners, mostly by choice. As we get older, we tend to earn more money, establish roots in a particular community and neighborhood, and start families. Consequently, the rate of homeownership increases with both age and income. However, single-family housing--which is preferred by most homeowners--is more costly in metropolitan areas where the demand for land for competing uses bids up land costs and results in higher density development. Metropolitan areas attract more renters for other reasons as well. For example, these areas are more likely to provide job opportunities for persons, particularly younger workers, willing to relocate from other areas. This creates a demand for rental housing for households who have not established an attachment to any particular area.

The overall homeownership rate for Virginia was 66.3 percent in 1990, a slight increase over the 1980 rate of 65.6 percent. Homeownership is lowest in higher density, urban regions and highest in rural areas. This is clearly reflected in the differences between homeownership rates for Planning District Commissions shown in Table 3. Homeownership rates in areas such as the Cumberland Plateau, Northern Neck, and the Middle Peninsula PDCs are around 80 percent. In contrast, homeownership rates in the Northern Virginia, Thomas Jefferson, Richmond Regional, and Hampton Roads PDCs are approximately 65 percent.

State and PDCs Homeownership Rate

%

Virginia			66.3
Lenowisco			75.8
Cumberland Plateau			79.3
Mount Rogers			76.3
New River Valley			64.2
Fifth			68.6
Central Shenandoah			69.8
Lord Fairfax			70.7
Northern Virginia			64.0
Rappahannock-Rapidan			72.8
Thomas Jefferson			63.5
Central Virginia			72.9
West Piedmont			73.5
Southside			73.4
Piedmont			76.2
Richmond Regional			65.3
RADCO			74.9
Northern Neck			82.2
Middle Peninsula			81.2
Crater			64.6
Accomack-Northampton			72.2
Hampton Roads			58.9

Owner occupied units divided by total occupied units.

The impact of income on homeownership is readily apparent from Table 4. At incomes below $10,000 (in 1990 dollars), less than half of the households are homeowners. Homeownership at this income level often reflects elderly homeowners whose current incomes are much lower than when they bought their homes. Homeownership increases substantially at incomes above $35,000 and by $75,000 the homeownership rate reaches its maximum level around 90 percent.

Income 1980 1990

% %

< $10,000		47.6	45.8
$10,000 - $19,999		50.9	51.7
$20,000 - $34,999		63.0	59.5
$35,000 - $49,999		77.2	72.3
$50,000 - $74,999		84.6	81.8
> = $75,000		89.3	90.5

Owner occupied units divided by total occupied units for each income category. Homeownership rates for 1990 were estimated based on 1990 Public Use Microdata income categories.

Comparing homeownership rates for different time periods is difficult because of the effects of inflation. The income categories given in Table 4 are in constant, 1990 dollar values. The homeownership rates for 1980 were estimated from income categories that did not correspond exactly to the 1990 dollar ranges given in Table 4 and some interpolation between categories had to be done. The homeownership rates for 1990 were estimated from the Public Use Microdata Sample for Virginia (this table was not yet available from other 1990 Census products). Because of these estimation procedures, minor differences in Table 4 of one or two percentage points should be ignored.

Between 1980 and 1990, homeownership declined the most for the middle income groups of $25,000 to $34,999, $35,000 to $49,999, and $50,000 to $74,999. The decline in homeownership among these incomes likely reflects a drop in affordability and shifts in household types away from husband-wife families. The overall stability of the state's homeownership rate from 1980 to 1990 was apparently the result of increased real incomes within the state which offset the declining ownership rate for the middle income categories.

Homeownership rates also vary substantially by age, although part of these differences is due to differences in incomes between younger and older households. The homeownership rate for 25 to 34 year olds was only 47.3 percent in 1990, but averaged 69.4 percent over the next ten year age category and was 78.1 percent for householders aged 65 and over. (See Table 5.) Between 1980 and 1990, the homeownership rate for younger households dropped from 51.8 percent to 47.3 percent for 25-34 year olds. It also declined for 35-44 year olds, from 72.6 to 69.4 percent. These declines reflect demographic shifts away from husband-wife households (which have higher homeownership rates) and toward single-parent families and non-families (which have lower homeownership rates). These trends reflect complex relationships between changes in life-styles, consumer preferences, incomes, and housing costs.

Age 1980 1990

% %

15 - 24		20.9	17.0
25 - 34		51.8	47.3
35 - 44		72.6	69.4
45 - 64		78.7	80.4
65 +		74.4	78.1

Owner occupied units divided by total occupied units for each age category.

Homeownership rates for younger households have far more variation across the state than for older households. (See Table 6.) Although this could be due to life-style choices that vary from area to area, it is probably also a reflection of the impact of housing costs on homeownership opportunities for younger households in certain areas. For householders aged 65 and over, the difference between the lowest and highest homeownership rates at the PDC level is 15 percentage points. However, for householders aged 25 to 34, this difference is 28 percentage points. The Thomas Jefferson, Crater, and Hampton Road PDCs have particularly low homeownership among 25 to 34 year old householders. These three areas also have exceptionally low homeownership rates for 35 to 44 year old householders. The Northern Virginia PDC had homeownership rates of 46.6 percent for 25-34 year old householders, 68.6 percent for 35-44 year old householders, and 73.1 percent for householders aged 65 and older. The latter was the lowest homeownership rate among all the PDCs. The rates for the other two age groups were only slightly lower than state averages.

Median house values varied substantially across the state, much more so than homeownership rates. (See Map 16.) The lowest median house value was $34,400 (in Lee County) and the highest was $231,000 (in Arlington County). The areas with the highest median house values ($100,000 and over) are mostly in Northern Virginia, with the addition of Albemarle County and a few counties around Williamsburg. The impact of suburban and exurban growth on house values is apparent in the high median values in and surrounding the Northern Virginia, Richmond, and Norfolk metropolitan areas. The areas with the lowest median house values (below $50,000) are mostly in the remote, mountainous portions of west and southwest Virginia and in the southern portion of the state.

State and PDCs Homeownership Rate

Age

25-35 35-44 45-64 65 and Older

% % % %

Virginia	47.3	69.4	80.4	78.1
Lenowisco	59.5	75.7	81.9	84.0
Cumberland Plateau	67.3	81.2	84.4	85.7
Mount Rogers	59.7	75.5	84.8	83.6
New River Valley	47.1	72.7	83.9	82.5
Fifth	48.8	70.0	80.8	77.4
Central Shenandoah	50.8	72.3	82.0	81.3
Lord Fairfax	51.8	70.8	81.3	81.8
Northern Virginia	46.6	68.6	79.4	73.1
Rappahannock-Rapidan	53.4	74.4	81.2	83.9
Thomas Jefferson	39.1	66.8	81.9	82.9
Central Virginia	55.4	74.8	84.0	79.6
West Piedmont	54.4	73.3	83.2	81.3
Southside	54.8	69.1	81.1	82.4
Piedmont	56.4	73.4	85.4	85.1
Richmond Regional	47.1	71.6	78.8	74.2
RADCO	60.9	79.2	86.3	82.7
Northern Neck	58.2	79.9	90.1	88.3
Middle Peninsula	67.5	81.0	88.4	88.2
Crater	38.9	63.8	78.7	77.2
Accomack-Northampton	49.4	64.1	80.9	83.6
Hampton Roads	40.8	62.8	77.2	73.8

Owner occupied units divided by total occupied units for each age category.

The geographic pattern for the percent change in median house values between 1980 and 1990 closely mirrors the pattern just described for 1990 median values: highest in the metropolitan and exurban areas of the Fairfax-Charlottesville-Richmond-Norfolk region and lowest in the southwest. (See Map 17.) Some rural areas toward the Chesapeake Bay (Northumberland, Lancaster, and Middlesex counties) also had rapid increases in house values between 1980 and 1990, ending with median values between $75,000 and $99,999 in 1990.

Given the differences in median values across the state, it is not surprising that some areas have very few houses with values less than $75,000 and other areas have a substantial supply of such modestly priced housing. (See Map 20.) The areas with the least amount of units with values less than $75,000 are in Northern Virginia (including Fredericksburg City and Spotsylvania County), Albemarle County, Williamsburg City, James City County, York County, Poquoson County, and Virginia Beach City. Communities adjacent to these have larger proportions of houses with values less than $75,000. In the half of the state west of a line drawn by the Norfolk, Richmond, and Charlottesville metropolitan areas, three-fourths or more of the houses in most of the communities have values less than $75,000.

In many of these same areas with large supplies of houses valued below $75,000, less than half of the homeowners had mortgages on their properties. (See Map 22.) In addition, only 50 to 75 percent of owners in many cities of the state had mortgages. This could reflect a combination of factors, such as elderly homeowners who have paid off their mortgages, low incomes making it difficult to qualify for loans, a lack of lenders in the local market, and restrictive lending practices. In general, the suburban communities of metropolitan areas have relatively high percentages of mortgaged owner-occupied housing units.

Monthly payments of principal and interest (which reflect the combination of prices and loan terms), along with utilities and taxes, constitute monthly homeowner costs. Median monthly costs for homeowners with a mortgage, as shown in Map 23, were below $500 in much of the area west of the Charlottesville-Richmond-Norfolk line. Median monthly costs throughout most of the metropolitan areas of Northern Virginia, Charlottesville, Richmond, and Norfolk were above $700. Median monthly ownership costs (for those with mortgages) were over $1,000 in Arlington, Fairfax, Fauquier, Loudoun, Prince William, and Stafford counties and in Alexandria, Fairfax, Falls Church, and Manassas cities. Median homeowner costs increased twofold or better throughout virtually the entire northern portion of the state and in much of the remainder of the Fairfax-Charlottesville-Richmond-Norfolk region. (See Map 24.) Again, areas to the west of the Charlottesville-Richmond-Norfolk line had substantially lower increases in median costs.

Median gross rents (rent plus utilities) followed the same pattern as shown for house values: substantially lower to the southwest of the Charlottesville-Richmond-Norfolk line and higher in and to the north of these areas. (See Maps 18 and 19.) Median gross rents were below $300 in most of southwest Virginia and along the southern border west of the Norfolk metropolitan area. Median gross rents in most of the communities in the metropolitan areas and a few exurban areas of the Fairfax-Charlottesville-Richmond-Norfolk region and in Frederick County were $450 and above. In Fairfax County, Loudoun County, Fairfax City, and Falls Church City median gross rents exceeded $800. Median gross rents also increased substantially (100 percent or more) between 1980 and 1990 in many of the same areas that had the highest rents by 1990.

Areas with proportionately larger supplies of low cost rental units (30 percent or more of rentals with gross rents below $250) were almost exclusively near the southern border and in the southwest. (See Map 21.) Buckingham and Northampton counties were the only exceptions. In other communities in the southwest and south, between 20 and 30 percent of rental units had gross rents below $250. Less than 10 percent of the rental market in most of the communities throughout the Fairfax-Charlottesville-Richmond-Norfolk region had gross rents this low. In several areas 5 percent or less of the rental market would be affordable to households with incomes less than $10,000 based on a standard of 30 percent of income for housing costs. These areas, along with the percent of renter households with incomes below $10,000, are: The counties of Albemarle (12.8%), Arlington (9.5%), Chesterfield (11.8%), Fairfax (7.0%), Fauquier (10.0%), Henrico (13.1%), King George (8.8%), Loudoun (7.4%), Prince William (6.1%), Spotsylvania (8.9%), Stafford (12.1%), and York (10.7%) and the cities of Alexandria (8.7%), Colonial Heights (13.4%), Fairfax (6.4%), Falls Church (11.6%), Manassas (5.9%), Poquoson (8.2%), Salem (14.2%), and Virginia Beach (10.0%). Such affordable rental units are proportionately more available in some of the exurban communities of the Fairfax-Charlottesville-Richmond-Norfolk region, but the overall number of units is small and, of course, much more remote.

The burden of high housing costs affects low-income households most seriously, forcing many of these households to devote more of their income to housing (and less to other necessities) than is widely accepted as advisable. The impact of housing costs on incomes is readily apparent in Figures 4 and 5, which show the percent of owners and renters, respectively, with housing costs in excess of 30 percent of household income. Sixty percent of owners and 80 percent of renters with incomes below $10,000 have housing costs which violate this affordability standard. The proportion of owners with unaffordable housing costs is cut in half (to 31.2 percent) for those with incomes of $10,000 to $19,999, but any significant additional reduction does not occur until incomes reach $35,000. At incomes of $10,000 to $19,999, renters are still substantially more likely to be burdened with higher housing costs than owners, with 63.5 percent exceeding the affordability standard. The proportions of renters at higher incomes with excessive housing costs are equal to or below those for owners.

The long-term benefit of homeownership is clear when comparing the percent of elderly households violating the cost-burden ratio, given in Figure 6. For elderly renters, 56.0 percent have housing costs in excess of 30 percent of income. In contrast, only 17.8 percent of elderly homeowners (who often have paid off their mortgage or have very little housing related debt) have to devote this much income to housing. By age 65, most householders who are still renters are those with very low life-time incomes. Renters under age 65, particularly younger renters, are influenced more by life-style choices and include a larger range of incomes. Consequently, a much lower percentage of these renters (36.6 percent) have housing costs in excess of 30 percent of household income. In contrast, the proportion of cost-burdened owners is slightly higher for those under age 65 (21.2 percent) than for those aged 65 and over (17.8 percent).

Low percentages of owners (less than 15 percent) pay in excess of 30 percent of income for housing costs in the western half of the state. (See Map 25.) In contrast, 20 percent or more of the homeowners in most of the communities in the Norfolk metropolitan area and in Northern Virginia are housing-cost burdened. Throughout much of the state, the ratio of owner costs to incomes declined during the 1980s. (See Map 26.) However, this ratio increased by 10 percent or more in Arlington, Dickenson, Fairfax, Fauquier, Gloucester, Greene, King George, Loudoun, Prince William, and Rappahannock counties and in Chesapeake, Manassas, Manassas Park, Virginia Beach, and Waynesboro cities.

When renter costs are considered relative to incomes, southwest Virginia emerges as one of the least affordable rental areas of the state, with 45 percent or more of renters paying in excess of 30 percent of income for rent. (See Map 27.) Although rents are relatively low, incomes are obviously even lower than elsewhere. Throughout many rural areas of the state, from 25 to 45 percent of renters had housing costs that exceeded the affordability standard. A much more varied pattern of changes in rent to income ratios also exists. (See Map 28.) The largest concentration of increases was in southwest Virginia, although several areas in the Fairfax-Charlottesville-Richmond-Norfolk region also witnessed increases in renter cost burdens. The pattern of increases in renter costs burdens compared with that shown for increases in gross rents suggests that rental affordability problems are much more a reflection of income patterns than rents. Areas marked by low rents and below average increases in rents had incomes which were still inadequate relative to costs.

Housing Quality

One of the primary measures of the physical quality of housing is the percentage of units that lack complete plumbing facilities for the exclusive use of the occupants of the unit. Some of these units lack running water; others have shared bathroom facilities, as in many boarding houses. In 1940, 64.9 percent of the housing units in Virginia lacked complete plumbing. (See Figure 7.) Some improvement was made during the 1940s, but major advances in housing quality occurred after 1950. Between 1950 and 1960, the percentage of units without complete plumbing declined from 28.2 to 12.9 percent and then dropped to 4.7 percent by 1980 and to 1.9 percent by 1990. As this percentage approaches zero, progress becomes more difficult in eliminating all of the residual units without plumbing.

Most of the remaining units that lack complete plumbing are in remote, rural areas and are occupied by poor households with very limited housing options. The elderly poor are particularly reluctant to upgrade their housing by moving. The significant improvements in this indicator of housing quality made in earlier decades were largely the result of the abandonment of obsolete units or upgrading by occupants who could afford improvements. Further improvements will likely rely on specialized housing programs, which provide assistance in upgrading housing for low-income occupants.

Units lacking complete plumbing are primarily found in the rural regions of the state. (See Map 29.) The highest percentages (6.0 to 13.4 percent) are found in rural areas in the eastern, southern, and western sections of the state, and a line of counties from Rockbridge to Amelia and in Charles City County. Toward northern Virginia, only Rappahannock and Madison counties have such high percentages of units lacking complete plumbing.

Other measures of housing quality collected in the 1990 Census include units lacking complete kitchens for the exclusive use of the occupants, the type of water and waste water system used by the unit, the number of rooms per unit, and the number of people per room (a reflection of overcrowding). Even fewer units lack complete kitchens (1.3 percent) than lack complete plumbing and the geographic pattern generally reflects the pattern for units lacking complete plumbing.

Seventy percent of all housing units are on sewer systems and 28 percent use septic tanks or cesspools. Only 2 percent have some other means of waste water disposal, most likely privies or direct discharge into a creek or stream. The communities with the highest percentages (6 to 18 percent) of units without sewer, septic, or cesspool waste water disposal are scattered throughout the state and are not confined to the mountain ranges. (See Map 30.)

Three of four units have a public or private source of water and most of the rest use drilled wells. (See Table 7.) About 4 percent of the housing units use dug wells and slightly over two percent have some other water source, usually a creek or rain water collected in cisterns or barrels. Fewer than 5 percent of the housing units outside the mountain ranges of the state obtain water from some other source than a public or private water system. (See Map 31.) The mountainous terrain of the western half of the state makes water system service more difficult. In most areas outside of cities, from 5 to 40 percent of the units are without water system service. The highest percentages (20 percent and over) are found in Bath, Bland, Floyd, Grayson, Highland, Lee, Rockbridge, Russell, and Scott counties.

Source of Water Units Percent

Public system/private company		1,900,436	76.1
Individual drilled well		444,953	17.8
Individual dug well		94,284	3.8
Some other source		56,661	2.3

Universe: Total Housing Units

Most wells, particularly drilled wells, provide a safe source of water and lack of a water system is not necessarily a housing problem. A few houses get water from creeks, cisterns, or rain barrels, which are clearly problematic sources for residential water use. This is a fairly prevalent problem in the mountains, particularly in the Lenowisco, Cumberland Plateau, Mount Rogers, and Central Shenandoah PDCs, where 10 percent or more of the housing units rely on potentially hazardous water sources. (See Table 8.)

State and PDCs Water Sources

%

Dug Well Drilled Well

Virginia		17.8	3.8
Lenowisco		24.1	3.4
Cumberland Plateau		40.5	3.6
Mount Rogers		30.6	3.3
New River Valley		23.0	1.7
Fifth		15.0	1.1
Central Shenandoah		27.6	2.4
Lord Fairfax		37.5	2.9
Northern Virginia		5.0	0.5
Rappahannock-Rapidan		54.7	6.1
Thomas Jefferson		34.8	6.6
Central Virginia		35.5	4.0
West Piedmont		41.7	5.5
Southside		51.3	15.2
Piedmont		57.3	16.5
Richmond Regional		8.7	5.1
RADCO		27.1	13.1
Northern Neck		36.5	27.8
Middle Peninsula		61.1	22.6
Crater		18.7	9.4
Accomack-Northampton		57.8	10.7
Hampton Roads		5.0	1.1

Universe: Total Housing Units

After a slight decline from 1940 to 1950, the median size of housing units in Virginia has progressively increased by about 0.3 rooms every ten years. (See Figure 8.) By 1990 the median number of rooms was 5.8.

Household Characteristics

Although houses have gotten larger, the number of persons per household has decreased, allowing even more space per person. (See Figures 9 and 10.) The number of persons per household has declined substantially over the past five decades, from 4.3 in 1940 to 2.6 in 1990. The number of housing units with 0.75 or fewer persons per room has increased dramatically over time. Between 1980 and 1990, the number of units with 0.75 and 1.00 persons per room declined sharply. During the same period, there were declines in the number of units with 1.01 to 1.50 persons per room (considered to be overcrowded) and those with more than 1.50 persons per room (severely overcrowded), but the drop was not as rapid as that for units with 0.75 to 1.00 persons per room.

Overcrowding is now relatively rare. Fewer than 3 percent of the households in the state have 1.01 or more persons per room. Although there is little variation in this percent across the state, the areas with the highest proportions of overcrowded households are the Northern Virginia, Southside, and Piedmont PDCs. (See Table 9.) Doubling-up of families is another indication of overcrowding. Families living with other relatives or with friends are referred to as "sub-families". About 4 percent of the families in the state (63,158 families) were doubled-up in 1990. Most of these subfamilies included children under the age of 18 years (54,938 families). The number of subfamilies increased by 89.4 percent between 1980 and 1990. Subfamilies were least prevalent in the western portion of the state. (See Map 4.) Between 5 and 12 percent of the families were doubled-up in many of the other areas of the state, including cities such as Richmond, Norfolk, and Portsmouth; suburban communities such as Fauquier County, and many rural communities such as Mecklenburg County.

State and PDCs 1.01 or More Persons Per Room

%

Virginia			2.8
Lenowisco			2.6
Cumberland Plateau			2.3
Mount Rogers			1.7
New River Valley			1.7
Fifth			1.4
Central Shenandoah			1.6
Lord Fairfax			2.1
Northern Virginia			3.7
Rappahannock-Rapidan			2.8
Thomas Jefferson			2.7
Central Virginia			1.9
West Piedmont			2.7
Southside			3.7
Piedmont			4.3
Richmond Regional			1.9
RADCO			2.7
Northern Neck			3.0
Middle Peninsula			2.5
Crater			3.5
Accomack-Northampton			3.5
Hampton Roads			3.3

Universe: Occupied Units

One of the most significant changes in housing markets throughout the United States has been the shift in households away from husband-wife families and toward non-family households. (A non-family household is a person living alone or two-or-more unrelated persons living together.) The decline in the percentage of husband-wife families and the increase in non-families in Virginia is shown in Figure 11. Husband-wife families ranged between 72 and 79 percent of all households from 1940 to 1970. During the next two decades, however, their proportion of the total dropped to 62 percent in 1980 and then 57 percent in 1990. The shift was mostly to non-family households, which went from 8 percent in 1940 to 29 percent in 1990, rather than to single-parent households. Contrary to the popular perception that single-parent families are more prevalent, the proportion of single-parent families in Virginia was lower in 1990 (14.3 percent) than in 1940 (15.8 percent) and only slightly above the level in 1970 (11.6 percent). Although a greater proportion of families are single-parent, this is due to the decline in husband-wife families rather than an increase in single-parent families as a percent of all households.

Throughout most of the state, between 60 and 70 percent of all households are married-couple families, 10 to 15 percent are other families, and 20 to 25 percent are non-families. (See Maps 1, 2, and 3.) Married-couple families are proportionately less likely (and non-family households more likely) in cities and in the state's larger university communities. A cluster of counties in the southern portion of the state and another toward the Chesapeake Bay have exceptionally low percentages of married-couple families and high percentages of other families (mostly single-parent families) and non-family households.

Another important demographic characteristic associated with housing characteristics is race. Table 10 presents a variety of housing characteristics according to the race or ethnicity (in the case of Hispanics) of the householder. Blacks, Hispanics, Asians, and other races than White have substantially lower homeownership rates than do Whites. Hispanics have the lowest homeownership rate (40 percent) among "minority" households and Asians have the highest rate (60 percent). These two groups have substantially lower homeownership rates than for Whites despite having higher median household incomes. The lower median household income for Blacks provides a partial explanation for having lower homeownership rates than Whites.

Black householders have the highest rates for a variety of problematic housing conditions. They have the highest incidences of living in units that lack complete plumbing or complete kitchens; units that do not have sewer, septic tank, or cesspool waste-water systems; or units with subfamilies. Blacks have the second highest incidence for units with dug wells or rely on cisterns or creeks for water. Although the presence of sub-families is higher among all minorities than among Whites, this might reflect cultural differences rather than a housing problem. Otherwise, minorities other than Blacks have housing conditions which are generally comparable to those for Whites.

Race

Conditions White* Black* Hispanic Asian* Other*

% Own	70.7	49.1	40.4	60.0	56.2
% Lack Plumbing	1.1	3.9	1.3	0.5	1.4
% Lack Kitchens	0.6	2.6	1.3	0.4	1.9
% Dug Well or "Other"	5.7	6.8	1.1	0.5	6.9
% "Other" Sewage	1.4	3.5	1.0	0.9	1.2
% Sub-families	2.3	9.5	5.1	6.0	6.0
Median Household Income	$35,000	$21,375	$36,000	$41,867	$31,070
Median Value	$85,000	$52,500	$112,500	$162,500	$85,000
Median Gross Rent	$496	$393	$655	$625	$473

*Other than Hispanic

Universe: Weighted occupied housing unit record of the Public Use Microdata Sample

Probably no other factor is more important in determining housing conditions than income and a discussion of income measures provides a good conclusion to this review of housing patterns in Virginia. Median household income in 1990 (based on 1989 income) was $33,328 and median family income, which is normally higher because it excludes persons living alone and other non-family households, was $38,213.

Median incomes have increased substantially in Virginia over the past two decades, as shown in Figure 12. The highest incomes are found in a stretch of communities from Frederick County to Albemarle County to Virginia Beach, and in Roanoke, Bedford, and Botetourt counties (See Map 5.) The lowest incomes are in southwest Virginia, in the Southside and Piedmont areas, and in Highland County, Northampton County, and Petersburg City. The areas where median household income increased the most between 1980 and 1990 were in the suburban and exurban communities around Northern Virginia and Charlottesville (particularly in the Rappahannock-Rapidan PDC), and a cluster of communities in the Northern Neck and Middle Peninsula PDCs. The largest cluster of communities with the least change in median household income was in southwest Virginia. In some of these communities (for example, Buchanan, Dickenson, Lee, Russell, and Tazewell counties) income growth was extremely depressed. Median household incomes also increased slowly in many cities of the state.

Extreme low incomes, as reflected by the percentage of persons below poverty, are often accompanied by extreme housing conditions. One-in-ten persons in Virginia was below the poverty level in 1990. The poverty rates for children (under 18 years old) and the elderly (65 years and older) were somewhat higher (13.3 percent and 14.1 percent, respectively). Maps 7, 8, and 9 provide the geographic distribution for the overall poverty rate and for the poverty rates for children and the elderly. The southern and southwestern portions of the state, along with many cities, are the most severely affected by poverty conditions. Poverty among children and the elderly is particularly prevalent in these areas.

Documentation

The data for The Virginia Housing Atlas: Housing Trends and Patterns to 1990 were compiled from Decennial Census Bureau publications and magnetic tapes. The data for the text, tables, figures, maps and profiles came from assorted Census publications from 1940 to 1980 and the 1990 Census of Population and Housing, Summary Tape File 1 (STF1A) and Summary Tape File 3 (STF3A) computer tapes. Whenever possible, the STF1a tape was used since it provides 100-percent data, while the STF3a tape provides sample data. The publications used for trend figures were: the Sixteenth Census of the United States: 1940 (Housing, Vol. II, General Characteristics, Part 5: Rhode Island - Wyoming); the Census of Housing: 1950 (Vol. I, General Characteristics, Part 6, Texas - Wyoming); the 1960 Census of Housing (Vol. I, States and Small Areas, Part 8, Texas - Wyoming); the Census of Population: 1960. (Vol. 1); the 1970 Census of Housing (Detailed Housing Characteristics, Virginia); the 1970 Census of Population (Vol. 1, General Population Characteristics, Virginia); the 1980 Census of Housing (Vol. 1, Characteristics of Housing Units, Chapter A, General Housing Characteristics, Part 48, Virginia, HC80-1-A48); the 1980 Census of Population (Vol. 1, General Population Characteristics, Virginia); and, the 1990 Census of Population and Housing, Summary Population and Housing Characteristics, Virginia (1990 CPH-1-48). Table 10, Race and Housing Conditions (Percents and Medians), was created based on the Census of Population and Housing, 1990: Public Use Microdata Samples [machine-readable data files], Virginia 5% sample. The results may vary from published Census documents.

The 1990 maps were created using data from the 1990 Census of Population and Housing, Summary Tape File 1 (STF1A) and Summary Tape File 3 (STF3A) computer tapes. In addition, maps showing changes from 1980 to 1990 used data from the 1980 Census of Population and Housing, Summary Tape File 1 (STF1A) and Summary Tape File 4 (STF4B) computer tapes and the 1980 Census of Housing (Vol. 1, Characteristics of Housing Units, Chapter A, General Housing Characteristics, Part 48, Virginia, HC80-1-A48). The maps showing Virginia counties and independent cities and the map showing Virginia Planning District Commissions (PDCs) were produced using Atlas GIS software on a personal computer. In all instances no shading represents "less" and a progression of density of shading represents "more". The data for the variables used in creating the maps are provided for each jurisdiction in the Data section at the end of this publication.

The data for the profiles of the state of Virginia, the Planning District Commissions, and each county and independent city jurisdiction came from the 1990 Census of Population and Housing, Summary Tape File 1 (STF1A) and Summary Tape File 3 (STF3A) computer tapes, except for the number of square miles, which is from the 1990 Census of Population and Housing, Summary Population and Housing Characteristics, Virginia (1990 CPH-1-48).

The source and the universe for each of the 31 maps included in this publication are provided below.

Map Number and Title Source Universe

1: Percent of Households by Household Type: Married Couples STF1a* Households

2: Percent of Households by Household Type: Other Family STF1a Households

3: Percent of Households by Household Type: Non-Family STF1a Households

4: Ratio of Subfamilies to Families STF3a Subfamilies/Families

5: Median Household Income STF3a Households

6: Percent Change in Median Household Income, l980-l990 STF3a & STF4b,1980 Households

Cont.

Map Number and Title Source Universe

7: Percent of Persons Below Poverty STF3a

Persons for whom poverty status is determined

8: Percent of Persons Under l8 Below Poverty STF3a

Persons for whom poverty status is determined

9: Percent of Persons 65 and Older Below Poverty STF3a

Persons for whom poverty status is determined

l0: Percent Increase in Housing Units, 1980-1990 STF1a & STF1a, 1980 Housing units

11: Median Year Structure Built STF3a Housing units

12: Percent of Units Built Between 1980 and March, 1990 STF3a Housing units

13: Percent of Units Built Pre-1940 STF3a Housing units

14: Mobile Homes as a Percent of Occupied Units STF1a Occupied units

15: Percent Change in Mobile Homes, 1980-1990 STF1a & STF1a, 1980 Occupied units

16: Median House Value, 1990 STF1a Housing units

17: Percent Change in Median Value, 1980-1990 STF1a & STF1a, 1980 Housing units

18: Median Gross Rent STF3a

Specified renter-occupied housing units paying cash rent

19: Percent Change in Median Gross Rent, 1980-1990 STF3a & STF4b,1980

Specified renter-occupied housing units paying cash rent

20: Percent of Owner Units with Values Less than $75,000 STF1a

Specified owner-occupied

21: Percent of Rental Units with Gross Rents below $250 STF3a

Specified renter-occupied (denominator includes no cash rent)

22: Percent of Owners with a Mortgage STF3a

Specified owner-occupied (not computed eliminated from the denominator)

23: Median Monthly Costs for Homeowners with a Mortgage STF3a

Specified owner-occupied

24: Percent Change in Median Homeowner Cost, 1980-1990 STF3a & STF4b,1980

Specified owner-occupied

25: Percent of Owners Paying 30% or More for Housing Costs STF3a

Specified owner-occupied (not computed eliminated from the denominator)

26: Percent Change in Owner Cost to Income Ratio, 1980-1990 STF3a & STF4b,1980

Specified owner-occupied

27: Percent of Renters Paying 30% or More for Housing Costs STF3a

Specified renter-occupied (not computed eliminated from the denominator)

28: Percent Change in Renter Cost to Income Ratio, 1980-1990 STF3a & STF4b,1980

Specified renter-occupied

29: Percent of Units Lacking Complete Plumbing STF3a Housing units

30: Percent of Units without Sewer, Septic, or Cesspool Waste Water STF3a Housing units

31: Percent of Units without Water System Service STF3a Housing units

*Assume a 1990 tape unless the date is otherwise specified

The source and universe for each of the profile sheet variables are provided below.

Profile Variables Source Universe

Population STF1a* Persons

Area CPH-1-48

Housing Unit Characteristics

Units in Structure STF1a Housing units

Year Round Status STF1a

% Seasonal/Migratory (Seasonal vacant+migrant vacant)/Housing units

Total Housing Units Housing units

Year Structure Built STF3a Housing units

Vacancy Status STF1a

Owner Vacancy Rate (Owner occupied+vacant for sale)/Owner occupied

Renter Vacancy Rate (Renter occupied+vacant for rent)/Renter occupied

Financial Characteristics

Gross Rent STF3a Specified renter-occupied

Housing Costs STF3a Specified owner-occupied

Value STF1a Specified owner-occupied

Cost Burden STF3a

Owners Specified owner-occupied

Renters Specified renter-occupied

Tenure

Overall Homeownership Rate STF1a Owner occupied/Total occupied

Homeownership Rate

by Age STF3a Owner occupied by age/Total occupied by age

Housing Quality

Source of Water STF3a Housing units

Condition

Boarded Up STF1a Vacant boarded up/Housing units

Waste Water STF3a Housing units

Facilities

Lacking Complete Plumbing STF3a Housing units

Lacking Complete Kitchen STF3a Housing units

Cont.

Profile Variables Source Universe

Household Characteristics

Household Type STF1a Households

Household Size and Crowding

Population per Household STF1a Occupied housing units

1.01 Persons or More per Room STF1a Occupied housing units

Income and Poverty

Median Household Income STF3a Households

Median Family Income STF3a Families

Persons Below Poverty STF3a Persons for whom poverty status is determined

Subfamilies

Number of Subfamilies with

Children Under 18 STF3a Subfamilies

Ratio to Total Families STF3a Subfamilies with child <18/Families

*Assumed to be 1990 tape unless the date is otherwise specified

The state and the 136 county and independent city jurisdictions are geographic levels directly available from the tapes. The Planning District Commission's (PDCs), however, are not geographic levels available from the census tapes, so data were aggregated by computer using the jurisdiction level data of the PDC members. Since medians cannot be aggregated, interpolation was used to estimate the medians for the PDCs from grouped data. In addition, please note that Chesterfield County, which is in both the Crater and Richmond Regional PDC, is included in the Richmond Regional PDC tabulations and not included in the Crater PDC tabulations in this publication. The following list of PDCs and the jurisdictions included within each PDC is provided for reference.

1. Lenowisco ( 703-431-2206): Lee, Wise, Scott counties and the city of Norton

2. Cumberland Plateau (703-889-1778): Buchanan, Dickenson, Russell, and Tazewell counties

3. Mount Rogers (703-783-5103): Bland, Carroll, Grayson, Smyth, Washington, Wythe counties and the cities of Bristol and Galax

4. New River Valley (703-639-9313): Floyd, Giles, Montgomery, Pulaski counties and the city of Radford

5. Fifth ( 703-343-4417): Alleghany, Botetourt, Craig, Roanoke counties and cities of Clifton Forge, Covington, Roanoke, and Salem

6. Central Shenandoah (703-885-5174): Bath, Rockbridge, Rockingham, Highland, Augusta counties and cities of Buena Vista, Lexington, Staunton, Waynesboro and Harrisonburg

7. Lord Fairfax (703-636-8800): Clarke, Frederick, Page, Shenandoah, Warren counties and the city of Winchester

8. Northern Virginia ( 703-642-0700): Arlington, Fairfax, Loudoun, Prince William counties and cities of Alexandria, Fairfax, Falls Church, Manassas and Manassas Park

9. Rappahannock-Rapidan (703-829-7450): Culpeper, Fauquier, Madison, Orange, and Rappahannock counties

10. Thomas Jefferson ( 804-972-1720): Albemarle, Fluvanna, Greene, Louisa, Nelson counties and the city of Charlottesville

11. Central Virginia (804-845-3491): Amherst, Appomattox, Bedford, Campbell counties and cities of Bedford and Lynchburg

12. West Piedmont (703-638-3987): Franklin, Henry, Patrick, Pittsylvania counties and cities of Danville and Martinsville

13. Southside (804-447-7101): Brunswick, Halifax, Mecklenburg counties and the city of South Boston

14. Piedmont (804-392-6104): Amelia, Buckingham, Charlotte, Cumberland, Lunenburg, Nottoway, and Prince Edward counties

15. Richmond Regional (804-358-3684): Charles City, Chesterfield, Goochland, Hanover, Henrico, New Kent, Powhatan counties and the city of Richmond

16. RADCO (703-373-2890): Caroline, King George, Spotsylvania, Stafford counties and the city of Fredericksburg

17. Northern Neck (804-529-7400): Lancaster, Northumberland, Richmond, and Westmoreland counties

18. Middle Peninsula (804-758-2312): Essex, Gloucester, King and Queen, King William, Mathews, and Middlesex counties

19. Crater ( 804-861-1666): Chesterfield*, Dinwiddie, Greensville, Prince George, Surry, Sussex counties and cities of Colonial Heights, Emporia, Hopewell, Petersburg

22.**Accomack-Northampton (804-787-2936): Accomack and Northampton counties

23. Hampton Roads (804-420-8300): Isle of Wight, James City, Southampton, York counties and cities of Chesapeake, Hampton, Newport News, Norfolk, Poquoson, Portsmouth, Virginia Beach, Williamsburg, Suffolk, Franklin

*Included in Richmond Regional PDC for all Atlas tabulations.

**PDCs 20 and 21 were combined as PDC 22.