Page images
PDF
EPUB

PART TWO

HOUSING SUPPLY AND NEED

Introduction

Very little was known about the characteristics of housing, real estate and construction until the depression of the early 1930's brought about widespread interest in this phase of the economy. The collapse of real estate values, both residential and nonresidential, and the apparently moribund condition of the construction industry suddenly awakened private individuals and public officials to the fact that not only did we lack well-formulated policies to be applied in dealing with these important segments of the economic structure, but that we did not even possess the basic data, analyses and interpretations, necessary for the development of such policies.

Authoritative estimates of the current relative magnitudes of the value of real estate and the total wealth of the United States are not available. However, in 1938, the most recent year for which estimates exist, the value of real property and improvements represented over half the national wealth. This relationship appears to have held in the late 1930's and from the turn of the century through 1922. Large gaps in the data make it impossible to state conclusively that this relationship has been fairly constant, but from the data available this would appear to be the case. Little information is available which shows the importance of the value of nonfarm2 residential real estate in relation to all real property and improvements, but an idea of the general magnitudes can be gleaned from the fact that in 1930 about 45 percent of all nonfarm real estate represented residential property.

Since housing is a durable good the characteristics of the total inventory loom even more important than the year to year increments, significant as they are in the functioning of the economy.

1 National Industrial Conference Board, The Economic Almanac for 1946-47, New York, 1946. See also David L. Wickens' Residential Real Estate, National Bureau of Economic Research, New York, 1941, pp. 1-2. 'Nonfarm as used herein, includes all urban and rural nonfarm areas. A small number of urban farm units are thus included except where otherwise noted.

Also because housing is an economic good that may last a lifetime, if the home is owned it is generally the largest single purchase made by a family, while the upkeep or rent of dwelling units constitutes a large share of the budget of most families.

Development of the Housing Census

In the light of the importance of housing both economically and socially, our knowledge of the housing supply in itself and in relation to requirements was scant until the 1940 housing census was conducted. Up to that time, census information relating to housing was gathered as a byproduct of the population count. As a result, on a national scale, little was known about the characteristics of the housing supply save the total number of occupied dwelling units and some break-downs in terms of tenure, color of occupants, and urban, rural, or nonfarm area. Not until 1930 were the value or rent and type of structure of dwelling units reported, and then only occupied units were covered. The way in which definitions of some of these concepts have been set forth in the Bureau of the Census' instructions to enumerators reduced the value of such information as was collected on these points in 1930. For example, in 1930, data showing type of structure are weak because of the exclusion of vacant units. If a multiple dwelling structure had several vacancies and was occupied by only one family it was enumerated as a onefamily dwelling.

Much more detailed information was collected in the real property inventories and surveys which were initiated as work relief projects during the thirties in many cities. The inventories represented complete enumeration while the surveys were conducted on a sample basis. Some were federally sponsored and others were locally sponsored, and for that reason there is some variation in the items included. In 1934, a comprehensive

survey was undertaken of 61 cities or metropolitan areas which were selected on a broadly distributed geographic basis with at least 1 city chosen in each State; but the results are best suited for use in local analyses. Because of the breadth of the survey, it has been claimed that the results have national significance, but there is some doubt as to whether, in the light of present knowledge of sampling techniques, the method of selection of areas covered could be supported scientifically, and the results used without exhaustive analysis and selection. Such a comprehensive study was undertaken by the National Bureau of Economic Research and the results were published in Residential Real Estate. Not only were the real property surveys and inventories drawn upon heavily in preparation of this volume, but many other sources were utilized as well, since the purpose of the study was to reveal the economic position of residential real estate in terms of values, rents, family incomes, financing and construction. This is one of the most important sources of general information relating to housing prior to 1940.

Two volumes which summarized the results of some of the inventories and surveys referred to previously are the Department of Commerce's Financial Survey of Urban Housing, 1937, and the Works Progress Administration's Urban Housing, 1938. Although this material is now old, much important information was collected as to type of structure, age and condition, number of stories, existence of basement and garage, value and mortgage status, facilities, characteristics of occupants and doubling in the localities where this research was conducted. The most comprehensive file of these studies may be found at the Federal Housing Administration library.

The Census of Housing in 1940 marked the coming of age of housing inventory statistics. Data were presented on housing and related items. for the country as a whole, urban and rural nonfarm areas, geographic regions, divisions, States, counties, metropolitan districts, cities, minor civil divisions, and tracts and blocks, thus making many comparisons possible for the first time. For those who wish to explore further the many facets of housing there is a bulletin called Housing, Index of Reports, published by the Bureau of the Census in 1945 which gives subject references to the volumes issued on the basis of the 1940 Census of Housing. Subject Guides to subsequent surveys and reports of the Census Bureau may be found in the periodic supplement it publishes.

Sample Surveys of Housing

Sample surveys of the national housing situation have been conducted from time to time by the Census Bureau in connection with its Monthly Report on the Labor Force. The first comprehensive national sample survey of housing published by the Census Bureau was made in October 1944. A second housing survey to which extensive reference is made in the subsequent chapters of this section was conducted in November 1945. The sample included over 20,000 households distributed in 68 areas, each consisting of one or more counties located in 42 States and the District of Columbia. Because of the nature of the sample, however, data for individual localities are not available. Only national totals have significance, and they must be considered in the light of the error due to sampling variability. Apparent small changes between 1940 and 1945 may not be meaningful in either the amount or direction of change. A more detailed statement of sampling error may be found in the census releases of the November 1945 survey.3

A third sample study was made in April 1947 when the Bureau of the Census conducted its most inclusive national survey of housing conditions and family characteristics. The results of this survey will be made available in the fall of 1947. The survey covered 148 areas. Because of the extended coverage of the sample, statistics for the Northeastern and North Central States, the South and the West will be published as well as for urban and rural areas of the United States as a whole. In addition statistics will be made available for the following 34 metropolitan areas:

Akron, Ohio. Allentown

Bethlehem

[merged small][merged small][merged small][merged small][ocr errors][merged small]

New Orleans, La. New York, N. Y.

[merged small][ocr errors][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small][merged small]

3 Characteristics of Occupied Dwelling Units, for the United States November, 1945, series H-46, No. 1, May 1946 and Vacancy in Dwelling Units in the United States: 1945, serics H-46, No. 2, June 1946.

The housing statistics obtained in April 1947 cover occupancy and vacancy, tenure, monthly rent, type of structure, number of rooms, state of repair, plumbing facilities and equipment included in the dwelling units. Data on characteristics of families, heads of households, and secondary families will be made available in considerable detail. Further information on characteristics and marital status of the population will likewise be published.

A more limited national survey showing housing intentions and living arrangements of veterans was conducted by the Census Bureau in June 1946. The results of this study are presented in National Housing Agency's Statistics Bulletin No. 7, A National Survey of Veterans' Housing Plans and Present Accommodations as of June 1946. This study presents data upon the quality and condition of housing occupied by veterans as well as an indication of the types and prices of housing sought by veterans in the market for housing.

On a local basis many housing surveys have been conducted since 1940 by both the Bureau of the Census and the Bureau of Labor Statistics. The earlier surveys were mainly of occupancy and vacancy but the more recent surveys cover the living arrangements, income, race, and housing intentions of veterans.

These various sample surveys of the housing situation have been designed primarily to reflect the condition of the housing inventory as of a given date rather than to measure net changes occurring between two closely adjacent points in time.

While it is possible to use these studies to measure the approximate order of magnitude of significant changes over time, errors due to sampling variability are such that the figures may not satisfactorily measure small changes. For example, the sample housing survey conducted by the Census Bureau in October 1944, estimated the number of occupied nonfarm dwelling units to be about 30,750,000 and the chances are about 19 out of 20 that the result which would have been obtained from a complete census using the same schedules, instructions, and enumerators would be between 30,050,000 and 31,450,000, a deviation of about 2 percent. The survey conducted in November 1945 using the same sample design in the same 68 sample areas as in October 1944 estimated the number of occupied nonfarm dwelling units to be about 31,300,000; the chances are about 19 out of 20 that the result which would have been obtained from a complete census is between 30,500,

000 and 32,100,000, a deviation of only about 3 percent. While a comparison of the sample survey results would indicate an increase of 550,000 occupied nonfarm dwelling units between the two studies the chances are 19 out of 20 that because of sampling variation the actual change might have been greater or less than the survey results by as much as 1,000,000 units. In other words, due to sampling, while there might be an error of only 2 or 3 percent in the level of a given survey figure there may easily be a relative variation of 100 percent or more on the observed change between the figures of the two surveys when that change has been relatively small.

As a further example, the survey conducted in November 1945 estimated the total number of nonfarm dwelling units to be about 33,340,000. The survey conducted in April 1947 will estimate the total number of dwelling units to be about 34,250,000. The chances are about 19 out of 20 that if there were no real change between the two dates in this case the sample estimates might differ up to 1,100,000. However, since the observed difference is about 900,000, a complete census on the two dates could have shown a slight decrease or an increase as large as 2,000,000.

When comparisons are made with the 1940 census data, on the other hand, the possible range of error in a net change figure is lessened, since sampling variation is not present in a complete enumeration. Moreover, when a longer period of time is involved, larger and hence more significant changes in level have generally occurred. For example, in 1940 the census showed 15,196,000 owner occupied units in the United States, while the 1945 sample survey showed about 20,009,000, an increase of 4,813,000 units. Since the chances are about 19 out of 20 that a complete enumeration in 1945 would have shown a figure not more than 850,000 greater or less than that obtained in the sample study, the indicated increase between 1940 and 1945 can be regarded as significant. Refinement of Data

As more experience was gained during the war in enumeration of dwelling units and their occupants, Census' definitions and field instructions to enumerators were refined. Although it would be advantageous from the point of view of strict comparability to have maintained identical definitions from the time of the 1940 census throughout all the surveys, both national and local, that were made, it is believed that the tightening of con

cepts represented improvement in the quality of the statistics collected. Particularly in the course of conducting vacancy and occupancy surveys did it become important to distinguish usable dwelling units from furnished rooms equipped for light housekeeping and from shacks and tenements that were virtually unfit for habitation. The 1940 census definition gave only general instructions to the enumerator as to how to distinguish dwelling units from rented rooms and made no attempt to screen out vacant units verging on the uninhabitable. The resulting complications were relatively unimportant until vacancies approached the vanishing point. In 1940, as the Census used the term, a dwelling unit consisted of the "living quarters occupied by or intended for occupancy by one household." Since a household was defined as a "family or any other group of persons living together, with common housekeeping arrangements, in the same living quarters," some difficulties were experienced by enumerators in determining the existence of a separate household and consequently whether the quarters visited actually constituted a separate dwelling unit. Thus, in some cases, furnished rooms were counted as dwelling units, while in other cases they were classified as rooms and their occupants as doubled families, lodgers, boarders, etc. It appears that, in the 1940 census, a number of furnished rooms were returned as separate units, thereby overstating the number of dwelling units.

Certain Concepts Refined

Since the definitions of dwelling unit and household are so closely related (a dwelling unit being the space occupied or intended for occupancy by one household), the refinement of both these concepts progressed simultaneously.

For the earlier surveys, the 1940 definitions of dwelling unit and household were used with several minor changes. The first major refinement was made by the Bureau of the Census in May 1944, when the following was included in the dwelling unit definition to assist the enumerators in distinguishing a dwelling unit from a room: In a structure where certain persons or groups of persons share some or all of the housekeeping facilities, each person or group of persons that meets one or more of the following conditions should be considered a separate household occupying a separate dwelling unit, "(1) each group has exclusive use of all rooms it occupies, (2) each group has exclusive use of all rooms it occupies except the bathroom,

(3) each group has exclusive use of all rooms it occupies except the kitchen, provided that food is prepared separately. If, however, food is shared and the groups eat together, all persons concerned are to be considered one household." This definition was used by the Census Bureau for the November 1945 housing supplement to the Monthly Report of the Labor Force.

Subsequently emphasis was shifted to the physical arrangement of a dwelling unit and the Census Bureau set forth the following criteria which are currently used to aid enumerators in distinguishing a dwelling unit: In general, a dwelling unit is a group of rooms or a single room occupied or intended for occupancy as separate living quarters by a person living alone, or by a family or other group of persons living together. A group of rooms is a dwelling unit if (1) it has separate cooking facilities which must be more substantial than a two-burner hot plate or an electrical appliance or (2) it has a separate entrance. A group of rooms is considered as having a separate entrance if no one has to pass through them to get to his own living quarters, and if the occupants of the group of rooms need not pass through any other living quarters to reach their rooms. single room is considered a dwelling unit provided that (1) it has separate cooking facilities, or (2) it has a separate entrance and a private bath, or (3) it has a separate entrance and is rented unfurnished.

Α

In 1940, all occupied dwelling units regardless of condition and all vacant dwelling units in habitable condition, were counted as dwelling units. In the dwelling unit vacancy and occupancy surveys conducted during the war years, all occupied and all vacant units, both habitable and not habitable, were included in the dwelling unit count. However, surveys made during the early war years classified as habitable vacancies only those vacant units which were in good condition or in need of only minor repairs. With the increased tightening-up of the supply of housing, it was felt that vacant units in need of major repairs should also be considered part of the housing market and in September 1942, habitable was defined to include units in good condition or in need of minor or major repairs. Only unfit

Excluded from the count of dwelling units were hotels for transient guests, missions, cheap one-night lodging houses, student dormitories, nurses' homes, education, religious or military institutions, penal institutions, homes for orphans and for the aged, blind, deaf, infirm and incurable, CCC camps, and Coast Guard stations or railroad or other construction camps in which workers live in common barracks. Government-owned dwelling units on United States military reservations-private living quarters for families of officers or employees, and barracks for enlisted men-have not been included in the housing data.

for use vacancies were excluded from the habitable group. It soon became apparent, however, that, with this definition, habitable vacancies included a number of vacant units which were in need of major repairs and were located in a neighborhood where such units are not generally occupied by persons who would live there.

Accordingly, the following modification was made in the distinction between "habitable" and "not habitable" dwelling units. Classified as habitable were those unoccupied units which were in good condition or in need of minor repairs, or in need of major repairs and located in a neighborhood where similar units were occupied.

Those units which were judged totally unfit for use or in need of major repairs and located in neighborhoods where similar units were not occupied were considered as not habitable. It should be noted that under conditions of acute housing shortage, and at other times as well, the great bulk of dwelling units needing major repairs are being occupied in the neighborhoods where they occur. "Habitable" vacancies, therefore, continue to include dwelling units of seriously substandard character. A more detailed discussion of the problems and techniques of the surveys appears in chapter III of Part Two, Vacancy and Occupancy.

« PreviousContinue »