1960 housing units was later transcribed to a mark-sensing card along with the primary sampling unit and segment number. The data on the mark-sensing cards were then converted to tape and used in the production of the ratio estimates for new construction between 1950 and 1959. Edit of Punchcards After the completion of the punching, all punchcards for each schedule (H-7 and H-7a Inventory Changes Schedules and H-116 and H-116a transcription forms) were run through a multicolumn sorter and those cards with inconsistent or impossible entries were rejected. These error listings were examined by housing-subject specialists and were corrected by using the information in entries in other items on the listing as well as the corresponding schedules and the 1950 census and 1956 National Housing 'Inventory records. The punchcards were then corrected and reedited. Collation of Schedules with Master Segment Record Control counts consisting of the number of units by each schedule type had been entered on the H-7 and H-7a Inventory Changes Schedules for each area segment and subsequently punched on the H-7 and H-7a punchcards. By using the primary sampling unit and segment numbers, the H-7's and the H-7a's were collated with the Master Segment Record on a segment-by-segment basis on the computer, and the control counts for each schedule type on the H-7 or H-7a were transferred to the Master Segment Record for the segment. The actual number of the H-7 or H-7a punchcard was then compared with the control counts. If the numbers agreed, then the information for the segment on the Master Segment Record was merged with that for each H-7 or H-7a in the segment and put on another tape. If the counts did not agree, both the Master Segment Record for the segment and the H-7's or H-7a's were examined, and the necessary correction made. The H-8 Characteristics Schedules and the punchcards corresponding to H-116 and H-116a transcription forms were then collated with the Master Segment Record by the same procedure. The collation with the Master Segment Record and the H-141 cards which contained the total number of 1960 housing units found in the second visit to the segment (phase II) consisted in transferring this total to the Master Segment Record. Computer Edit of Components of Change The next step in the computer processing was to edit the components on the computer. A small portion of the computer edit specifications for the H-7 schedules is given below: the specifications for the H-7a, H-116, and H-116a records were roughly the same. Computer Edit of Characteristics The computer consistency edits of the 1959 items on the H-8 Characteristics Schedules were divided into two parts, one for the components transcribed from the H-7 or H-7a Inventory Changes Schedules and the other for the characteristics. The edits for the components were done first, and, for the most part, were identical to the computer edits for the components on the basic H-7 and H-7a schedules. The items on 1959 characteristics were edited for consistency in much the same manner as the corresponding 1960 housing items. The method of imputing the nonresponses (NA's) for characteristics was different from the one used for the 1960 census. Responses for the NA's in the 1959 items for condition, rooms, year built, and number of dwelling units in the structure were allocated in the case of "same" units by using the entries in the corresponding 1956 or 1950 items if they were available. For other units, these items, except for the one on the year the structure was built, were edited by using data from the preceding 1959 unit. For the item on the year the structure was built, when the 1956 or 1950 data were not available or applicable, and for the one on location of previous residence, the NA's were imputed on the basis of the 1956 National Housing Inventory distributions. Blanks in items on value, acreage and sales, contract rent, amounts spent for utilities, and plumbing were allocated directly from the preceding units. The use of data from the preceding unit in allocating where necessary for most of the 1959 items instead of the various imputation patterns used for the same 1960 housing items was due to the limited storage capacity of the Univac I computer on which most of the edits were made. Nonresponses in the 1959 items for value and for contract rent, and for the related utilities items used in the computation of gross rent, were allocated for the simple distributions of the 1959 characteristics. When these 1959 items were cross-tabulated with the corresponding 1950 or 1956 items for same units or with rent and value of previous units for recent movers, the NA's were retained to avoid unrealistic comparison of a reported amount with an imputed amount on a unit-by-unit basis. The edits for the 1950 and 1956 characteristics of dwelling units classified as "same" were, in general, similar to the edits for the 1959 items. NA's for items on condition, number of rooms, number of units in structure, year built, and plumbing were edited by using the entries in the corresponding 1959 items. The blanks in the remaining items were edited by using data in related items. The NA's were retained for items on value, contract rent, and gross rent. The data on 1950 characteristics transcribed to the H-116a transcription form and those on 1956 characteristics transcribed to the H-116 transcription form for units which were changed or lost to the 1950 or 1956 inventory were edited in the same manner as the 1950 or 1956 characteristics on the H-8 Characteristics Schedule, except that responses for certain NA's that could not be ascertained from related items were allocated from entries in the preceding unit. Recode of Characteristics Substantially more programing would have been required to produce final publication tables on characteristics from the high-speed printers like the 1960 census tabulations. Therefore, a special tabulation program which had been prepared for the Univac I computer for the 1956 National Housing Inventory was used. This program required that the data in each item be recoded in the computer by categories according to tabulation specifications. Restrictions for certain items were included in the recode program. For example, the entry on value of a dwelling unit was recoded into one of the predetermined categories only if the unit was owner-occupied, nonfarm, not a trailer, and was a one-unit structure on a one-unit property with no business on the property. Preparing Final Tables for Reproduction After the data had been tabulated, they were posted to work sheets and the percents and medians were computed. All the work was completely verified. Certain 1950 and 1956 historical data were also posted to the work sheets. The tables were given a preliminary internal consistency check by the clerks, then were reviewed for subject matter by housing-subject specialists. Sampling specialists prepared the variance tables for the publications and reviewed the summary findings. The tables, text, and variances were then typed; the tables were typed on preprinted table formats. After the typing, the tables were verified. The publication report was then assembled and sent to the printer. PUBLICATION PROGRAM Prior to detailed reports, press releases were issued containing the counts of the 1950-59 components of change with a brief analysis of their relation to the 1950 and 1959 housing inventory. The initial press release was issued in September 1961 for the Washington, D.C.-Md.-Va. SMSA. The last release, for the United States, was issued in March 1962. The final results of the Components of Change Survey were issued as series of reports which together constitute volume IV of the 1960 Census of Housing. In addition to the tables, each report contained a text including definitions and a summary of findings on the subject content, as well as a statement on sample design and sampling variability, and a map showing the boundaries of the area covered in the report. The first SMSA report for the series comprising volume IV of the 1960 Census of Housing, Components of Inventory Change, Part 1A, 1950-1959 Components, was published in December 1961. The last report of part 1A, for the United States and regions, was published in October 1962. Part 1A presents statistics on the counts and characteristics of the components of change in the housing inventory, 1950 to 1959, for the United States, by regions, and for the 17 SMSA's. Counts describing the source of the 1959 inventory as well as the disposition of the 1950 inventory are presented--i.e., new construction, conversions, and other additions, and mergers, demolitions, and other losses, and units that were the same in 1950 and 1959. Characteristics of the individual components, such as occupancy, structural and financial characteristics, and condition and plumbing facilities, are included. Additional selected household characteristics are shown for the total 1959 inventory and for new-construction units. The boundaries of the areas covered in the series were defined as of 1959 (see individual reports for detailed description of each area). Volume IV, Components of Inventory Change, Part 1B, Inventory Characteristics, presents statistics on the current and previous residences of recent movers (households that moved in 1958 and 1959), additional characteristics of units in new construction and "same" units for 1950-59, and characteristics of available vacant units. Reports were issued for the United States by regions and for the 17 SMSA's. The 18 reports were published from May to November 1962. Volume IV, Components of Inventory Change, Part 2, 1957-1959 Components, presents statistics, similar to those presented in part 1A, on the counts and characteristics of changes in the housing inventory after the December 1956 National Housing Inventory. Reports were issued for the United States by regions and for the nine metropolitan areas for which separate statistics were provided in the 1956 National Housing Inventory. The geographic boundaries of the areas presented in this series are the same as those used in the 1956 National Housing Inventory. The first report in this series appeared in December 1962 and the last in June 1963. During the processing of the Components of Change Survey data for publication, more data were tabulated than it was possible to include in the final reports. These unpublished data are available to users on a reimbursable basis upon arrangement with the Chief of the Housing Division. In general, they include additional tabulations relating to the total inventory, new construction, and recent movers. Chapter 3. RESIDENTIAL FINANCE SURVEY INTRODUCTION The Residential Finance Survey was conducted as a mail survey from late 1959 through the spring of 1960. The basic unit for the survey was the property. Questionnaires were mailed to a sample of property owners and to lenders who held mortgages on those properties. The sample was selected to represent a cross section of all nonfarm residential properties in the country. The principal objectives of the survey were: (1) to provide information on (a) the current mortgage status of nonfarm residential properties, (b) how these properties were acquired, (c) the number and amount of mortgages outstanding on these properties, (d) the characteristics of this indebtedness with respect to interest rates, terms of loans, Federal insurance or guaranty, extent of junior mortgages, and types of lenders; and (2) to relate the characteristics of the mortgages to the characteristics of the properties and, in the case of owner-occupied properties, to the characteristics of the occupants. As early as 1890, the Bureau (then the Census Division of the Department of Interior) collected detailed statistics on real estate mortgages. In 1920, a special mail survey was made on financing of nonfarm owner-occupied homes. In the 1940 Census of Housing, information was obtained on a limited number of mortgage finance items as part of the regular enumeration, but only for owner-occupied 1- to 4-family homes containing no business. In 1950, a separate sample survey on residential financing, very similar to the 1960 survey, was conducted by mail after the basic population and housing enumeration had been completed. In 1956, as part of the National Housing Inventory, a sample survey was made of owner-occupied properties having from 1 to 4 dwelling units, to obtain information on mortgage status and on the characteristics of the mortgages, properties, and owners. The 1959-60 Residential Finance Survey provided data on the financing of residential properties which was not covered in the 1960 Census of Housing. Information on certain of the property and structural characteristics collected in the Residential Finance Survey were also collected in the 1960 census, and the definitions were generally the same. However, there were several differences: 1. The basic unit of tabulation in the Residential Finance Survey was the "property" and not the "housing unit" used in the 1960 census or the "dwelling unit" used in the 1959 Components of Change Survey. 2. The Residential Finance Survey was restricted to nonfarm, privately-owned properties whereas in the 1960 census publicly-owned properties were included and only the data on value and rent were restricted to nonfarm properties. 3. The Residential Finance Survey was based on a subsample of dwelling units from a sample of land-area segments--the dwelling units enumerated in the Components of Change Survey to provide data on the 1959 characteristics of the components--supplemented by a sample of large rental properties. Data for the 1960 census were based on 100-percent coverage of all housing units for some items and on a systematic sample of housing. 4. The boundaries used in the Residential Finance Survey were as of June 8, 1959. In most cases these were the same as the boundaries used in the April 1960 census. PLANNING THE SURVEY Consultation In the summer of 1957, the Bureau conducted an informal survey among lenders, research personnel, educators, representatives of Government agencies, members of the Housing Advisory Committee, and members of the Bureau of the Budget's Committee on Residential Finance for the 1960 Census of Housing, to obtain information on the uses of the residential finance data provided in the past and to elicit suggestions concerning the relative importance of the several aspects of the subject. The general conclusions reached were the following: 1. The users of the 1950 census data on residential finance found them to be useful and important. 2. The varied interests of the users and the diverse ways in which the data were employed suggested that the 1960 census program should include collection of data from all participants in the mortgage-financing processes, lenders as well as owners, and owners of rental properties as well as homeowners. 3. Owner-occupied 1-dwelling-unit properties should be given priority in detailed crosstabulations. 4. Data on selected SMSA's should have high priority but no attempt should be made to supply data on areas smaller than SMSA's. To receive the benefit of advice from mortgage-finance experts outside the Bureau, a Residential Finance Technical Advisory Committee was established in 1958. An effort was made to include individuals who could contribute to the solution of the many technical problems associated with the program. Each committee member was requested to serve in the capacity of an individual expert rather than as a representative of the organization for which he worked. Some of the members had served in a similar capacity in connection with the 1950 Residential Finance Survey. (See appendix C for a list of the members and dates of meetings.) In addition to attending the committee meetings, the members gave advice in informal conferences with the Bureau of the Census staff members and in answer to inquiries by mail. Some committee members gave generously of their time by contacting lenders concerning the availability of information and the willingness of lenders to supply certain types of information to the Bureau. In the course of conferences with users, several gaps in existing knowledge of mortgage markets and home financing were discovered. A list of the items needed to help fill these gaps, together with items for which data were collected in 1950, were placed on a ballot and submitted to members of the Residential Finance Technical Advisory Committee in the fall of 1958 with the request that they indicate the priority of each item. This helped to determine which items should be omitted and which subject areas required additional information. Two subject areas became the basis of pretests--secondary mortgage market activity (i.e., purchase and sale of mortgages by financial institutions) and costs of maintenance and repairs for homeowners. Advice was also sought from housing and mortgage-finance specialists in other Government agencies who were in a position to evaluate the relative value of alternate proposals from the point of view of their agencies. Determination of Scope and Coverage Type of property included.--Although most Technical Advisory Committee members and other users agreed that priority should be given to single-family homeowner properties, it was decided that rental properties should also be included in the survey. It was the consensus that rental housing construction was likely to increase significantly in the future and that there was need for more information on the characteristics of the predominating conventional loans. Items of information collected.--Because changes had occurred in housing and mortgage markets after 1950, some 1950 items were dropped from the program and some new items were added. Among the 1950 items dropped from the 1960 survey were those on reasons for refinancing, relationship of owner to the head of the household, and occupation of owner. Because mortgage payments generally were paid monthly, it was decided not to publish data relating to frequency of payments, but, because this information was needed to compute mortgage payments, it was retained on the questionnaires. During the planning stage for the 1960 survey, most users agreed on the importance of collecting data on the prices paid for mortgages transferred from the original holders to lenders during recent years. Because concern was expressed concerning the willingness of lenders to report these data, Technical Advisory Committee members and Bureau personnel interviewed lenders in some principal cities including New York, Washington, Chicago, Philadelphia, and San Francisco. There was general agreement among those conducting the interviews that lenders would be willing to supply the data. These items were included in the 1960 residential finance questionnaires with the understanding that the resulting tabulations would be of an experimental nature and that they would not be included in the final publication. Another group of new items recommended for inclusion in 1960 by many users was concerned with owners' expenditures for maintenance, repair, and property improvement, since these items together with data on mortgage payments, taxes, and insurance would provide the complete picture on expenses of property ownership. The feasibility of collecting data on these items in a mail survey was tested in the Washington Pilot Study and the Philadelphia pretests described below. On the basis of the tests, it was decided that reasonable expenditure data could be collected in this type of survey but that the required additional testing and refining of concepts and wording of the questions could not be accomplished in the time prior to the date when the Residential Finance Survey schedules would have to be in final form. Moreover, more time would be needed to determine whether a mail survey would be the most desirable method of securing this type of data. (The Bureau subsequently undertook an extensive study of the problems, and initiated a quarterly Survey of Residential Alterations and Repairs, using the personal interview technique for homeowners and a mail survey for owners of rental properties.) |