Construction is one of the Nation's largest and most volatile industries. Construction statistics are used for (1) short-term analyses and projections of the impact of construction activity on the economy for fiscal and monetary policymaking, (2) long-term analyses of living conditions and economic growth for developing policies associated with obtaining adequate housing and capital investment, and (3) analyses of construction markets by contractors, labor unions and suppliers of construction materials.

Construction statistics are closely related to other areas covered in the Framework-housing, mortgage financing, and overall economic activity. These are discussed in the chapters on housing and community development, financial statistics, and the economic accounts. Hence, this chapter is brief, covering only those topics that do not fit conveniently into those related chapters. In order to make a full assessment of the needs for and recommendations regarding construction statistics it is essential to review these related chapters.

Statistical Programs

Statistics on construction activity are primarily collected and distributed by the Census Bureau. The Census construction statistics are collected monthly, quarterly, annually, and quinquennially.

Monthly data and annual summaries and benchmarks are published in a series of construction reports for (1) value of new construction put in place in current and constant dollars for all types of private and public construction, (2) housing starts and completions, and new one-family homes sold and for sale, (3) housing authorized by building permits and public contracts, and demolitions at the local area level, and (4) mobile home placements (conducted for the Department of Housing and Urban Development).

A quarterly sales price index for new one family houses sold is developed. Annual information is provided on the value of residential repairs, additions, alterations and replacements (data are collected quarterly but tabulated annually).

The quinquennial census of construction industries covers receipts, expenses, capital outlays, and so forth, of construction contractors, operative builders, land subdividers and developers. These data are published by industry and state.

The above data are based on survey information on building permits, contract awards, housing and other construction starts, housing sales and completions, construction expenditures, and operations of the construction industry. For the measures of construction put in place, these data are supplemented with statistical estimating procedures. The single family housing price index is based on a statistical regression analysis that adjusts for the changing structural characteristics of new housing. Construction cost indexes used to deflate other types of construction activity to constant dollars are obtained from other Federal agencies and private industry.

The Bureau of Labor Statistics (BLS) conducts periodic studies to determine the labor, materials and overhead costs and construction time required for various types of construction. These studies have been conducted roughly in five to ten-year cycles for single-family housing, private multi-family housing, public housing, college housing, nursing homes, hospitals, schools, highways, sewers, civil works, Federal office buildings and commercial office buildings. The BLS also collects monthly data on payrolls and employment in the construction industry and wholesale price indexes of construction materials.

The Department of Labor completed the initial phase of a new system for developing short, medium and long-term projections of employment on construction projects in 1978, and the full implementation of the program is planned for 1980. This Construction Manpower Demand System provides forecasts of the demand for approximately 30 construction craft occupations plus on-site technical and professional personnel for about 35 types of construction in local areas, States, Federal regions and the Nation. The forecasts are developed monthly for the first year of the projections, quarterly for the second year, and annually for the next three

years. Longer-term projections of ten and more years will be made for some energy-related construction sectors.

Uses and Users

Data on construction activity are used by Federal, State and local government agencies and private organizations for economic analysis, projections and policymaking. The Council of Economic Advisors; the Federal Reserve Board; the Departments of the Treasury, Housing and Urban Development, Labor, and Commerce; the Federal Home Loan Bank Board; and the Federal and Government National Mortgage Associations use the data in determining fiscal, monetary, employment, and regional economic development policies for the financing of construction to stabilize the economy, to assess the needs for new private and public capital investments, and to help in achieving the goal of providing adequate housing (see the chapter on housing and community development for housing goals). The House Banking, Finance and Urban Affairs and Senate Banking and Urban Affairs Committees, Joint Economic Committee, and Congressional Budget Office use these data in evaluating the need for new legislation in similar areas.

The Department of Labor uses the construction data and projections for providing a better match between job vacancies and workers in its employment service and training programs. Within the Department of Commerce, the Bureau of Economic Analysis uses the construction put in place figures in developing estimates of the gross national product and the construction labor and materials studies in preparing input-output tables (see economic accounts chapter). Also, the Commerce Department's Industry and Trade Administration prepares Construction Review, a monthly publication which summarizes a wide range of construction statistics.

In the private sector, construction statistics are used by manufacturers and distributors of construction products for company planning and marketing. Trade associations, labor unions, and financial organizations use these data to analyze market trends, negotiate wage agreements, advise on investments, and recommend legislation.

Advisory Committees

There are no continuing advisory groups charged primarily with reviewing the construction statistics. program. The advisory committees to the Census Bureau, for example, the American Statistical Association and the American Economic Association advisory committees and the Advisory Committee on

State and Local Area Statistics, occasionally offer suggestions to improve the construction statistics program. Special studies have been made for improving construction statistics. For example, the Cabinet Committee on Construction prepared such a report in 1970 on a wide range of topicscompensation, industrial relations, prices, finance, employment, existing structures, output and industry statistics. More recently, the Gross National Product Data Improvement Project recommended several improvements for measuring construction activity in current and constant dollars (see chapter on economic accounts).

Major Data Gaps and Recommendations

The most prominent data problems with construction statistics are the inadequacy of construction cost indexes, lack of measures of nonresidential maintenance and repair construction, weaknesses in the measures of single-family construction put in place and of residential repairs and alterations, and lack of information on dealer inventories of mobile homes and of capacity in the construction industry.

Construction Cost Indexes

It has long been recognized that the construction cost indexes used to deflate the bulk of construction expenditures are not representative of actual material, labor and overhead costs. Based on previous analyses, these indexes tend to overstate the increase in construction costs due to inflation and consequently lead to an understatement of the estimated volume of construction in constant dollars.

This problem exists for approximately 70% of all new construction-multifamily housing, industrial, commercial, educational and other nonresidential buildings, public utilities, and civil works. An essential weakness of these indexes is that they are based on pricing hypothetical structures of certain physical characteristics which were specified at some time in the past rather than of actual structures currently under construction. They also are inadequately adjusted for productivity change in the construction industry.

These indexes typically are provided by private organizations. Unfortunately, little information is available on the methodology for preparing the indexes. The Census Bureau presently is developing its own construction cost index for multifamily housing. It is planned for use around 1980. This will be consistent with the indexes for single family housing (Census Bureau) and highway construction

(Federal Highway Administration) which are satisfactory.

The Census Bureau and the Bureau of Labor Statistics should engage in a joint research and development project to provide more reliable construction cost indexes for nonresidential construction. These should be developed initially for nonresidential buildings, and subsequently for electric power plants, sewer and water facilities, and other nonbuilding structures.

In addition to the conceptual aspects of the research, an examination should be made of the appropriate vehicles to obtain the necesssary statistical information. For example, the BLS studies on labor and materials requirements for new construction and the Census data on characteristics of nonresidential buildings are primary sources of such information. These data sources should be reviewed for necessary improvements regarding the representativeness of the sampled projects, adequacy of the coverage of different types of construction, and frequency of the surveys. The BLS and Census data collection programs should also be integrated to avoid duplication.

Nonresidential Maintenance and Repair Construction

Substantial sums are spent on the maintenance and repair of nonresidential structures-roughly estimated at over $20 billion annually. In addition, these outlays are being bolstered by modifications to existing structures to conserve energy. However, there are no directly collected data on these expenditures.

Estimates of expenditures for nonresidential maintenance and repair construction should be collected in the quinquennial economic censuses. These data would be provided in the census of construction industries for the work done by construction contractors, and in the other economic censuses for work done under contract and by force account employees.

Single Family Residential Construction

Estimates of monthly construction put in place for single-family housing are developed indirectly by multiplying the number of housing units under construction from data on housing starts by unit construction costs from building permit data. The housing start data are adjusted for the estimated proportion of the building constructed each month and the unit cost estimates are adjusted for undervaluation in the building permit data and for architectural and engineering expenses. The adjustment factors for the building permit data and

architectural and engineering work are based on studies conducted in 1956, which now considerably out of date. Adjustment factors for the proportion of housing starts constructed each month are based on 1970-72 patterns; prior to 1978, they also reflected 1956 experience.

The Census Bureau should conduct a monthly survey that provides direct measures of the value put in place of single-family construction, analogous to the ongoing monthly survey for measuring multifamily residential construction.

Alternatively, the Census Bureau should update the adjustment factors used for the unit costs and construction time every 5 years. In developing these factors, consideration should be given to using the BLS studies on labor and materials requirements for new construction and the Census surveys of housing sales and completions.

Residential Repairs and Alterations

The Census Bureau's survey of expenditures for residential repairs, additions, alterations and replacements has relatively high sampling errors. In addition, due to the lack of quarterly tabulations of these data (collected quarterly but tabulated annually), estimates of the outlays for additions, alterations and replacements are extrapolated from annual levels for the quarterly gross national product estimates of fixed investment.

The Census Bureau should increase the sample of the survey of residential repairs, additions, alterations and replacements to significantly reduce the sampling error, and tabulate these data quarterly in time for use in preparing the gross national product estimate released 75 days after the quarter. Consideration also should be given to merging this survey with the Bureau of Labor Statistics' Continuing Consumer Expenditure Survey.

Dealer Inventories of Mobile Homes

A major consideration in short-term demand forecasts is the status of inventories. Thus, if inventories are low relative to sales, future output tends to be strengthened; concomitantly, if inventories are relatively high, this tends to depress future output.

Unsold new homes are one form of inventory, and for conventionally-built single-family houses, monthly data on new homes for sale are available. Comparable information of mobile home dealer stocks of new mobile homes is not provided. Consequently, current market information affecting

future trends for approximately one-fifth of new housing units is not available.

The Census Bureau should add monthly information on dealer stocks of mobile homes to its ongoing survey of mobile home placements.

Areas for Further Research

Capacity in the Construction Industry.-There presently are no statistics on capacity and capacity utilization in the construction industry. Such information would be useful for analyses of potential output, inflationary pressures, and capital investment requirements for the construction industry.

Because of difficult methodological problems associated with developing a statistical series on capacity and capacity utilization in the construction industry, an interagency task force should study pragmatic means for developing such measures. The study should evaluate the feasibility of developing such measures.

Inventory of Nonresidential Structures.-An inventory of the characteristics and location of

existing housing is developed decennially, and in less detail annually. However, there is no comparable information, even on a one-time basis, for the stock of nonresidential structures.

The provision of an inventory of nonresidential structures by type of structure, ownership, age, utilization, geographic area and other characteristics potentially could be used to develop projections of the demand for additional and replacement structures in local areas, for estimates of tangible national wealth, and possibly as a universe from which to conduct sample surveys of expenditures for repairs, replacements, additions and alterations to existing nonresidential structures.

Because of the considerable costs associated with developing an inventory of existing nonresidential structures, an interagency task force should examine the likely uses to which such information would in fact be put. The results of this study should then be evaluated to determine the merits for initially developing such an inventory and subsequently for periodically keeping it up to date.

Responsible Agencies

A well-rounded statistical program for criminal justice would provide comprehensive information on crime and its victims, on offenders, and about the administration of justice at the Federal, State, and local levels. At present, there are some 15 Federal agencies and thousands of State and local agencies involved in the collection, analysis, and use of criminal justice statistics. The primary collecting agencies at the Federal level are the National Criminal Justice Information and Statistics Service (NCJISS) of the Law Enforcement Assistance Administration (LEAA), the Uniform Crime Reporting Section of the Federal Bureau of Investigation, the Census Bureau, and the Administrative Office of the United States Courts. Currently most the major criminal justice data collection activities involve the LEAA in one way or another. For example, LEAA presently expends over $5 million annually on interagency agreements with the Census Bureau.

Although a great deal of criminal justice data are collected, little analysis is accomplished, despite LEAA's clear mandate to do so. Moreover, there is no activity within the Department of Justice itself, with the possible exception of the Office for the Improvement of Criminal Justice, which presently has a role of coordinating and analyzing criminal justice statistics.

There are a number of other organizations within the Department of Justice which could be involved with criminal justice statistical activities, but are not. The most important of these are the Offices of the U.S. Attorneys, which are the focal point for the Federal criminal justice process from arrest through adjudication.

Users

A major problem concerning criminal justice in the United States today relates to basic public misunderstandings concerning both crime and its impact, as well as the efficacy and cost of the criminal justice process. Therefore, a major target group for criminal justice statistics is the general public, through the agency of the press. Because of the

constitutional reservation of many justice activities to the States, criminal justice decisionmaking is extraordinarily fragmented; the information reported by the press impacts directly on decisionmakers in city councils, State legislatures, and leadership personnel in criminal justice agencies at all levels of government. Thus, it is essential that basic misunderstandings concerning crime and the administration of justice be corrected, and our general knowledge in this area expanded.

The academic community also is another important user of criminal justice statistics. The last decade has seen the establishment of several specialized schools of criminal justice as well as the development of criminal justice programs in general institutions of higher learning. The data currently being produced are a valuable resource for these institutions. More importantly, behavioral scientists in general, and criminologists in particular, are using criminal justice statistics to develop and evaluate their theories about crime and justice. Legislative, executive, and judicial decisionmakers rely heavily on the advice of these academics.

Another key user of criminal justice statistics is the U.S. Congress. The Senate and House Judiciary Committees in particular use data to design legislation and programs. The Subcommittee on Criminal Laws and Procedures of the Senate Judiciary Committee, for example, used a great deal of statistical information on judicial outcomes and sentencing policies among the various Federal courts in the development of legislation to codify Federal criminal statutues. The Juvenile Delinquency Subcommittee continues to use information on juvenile institutions and juvenile courts in its development and monitoring of juvenile justice programs. The Subcommittee on Crime of the House Judiciary Committee uses information on urban crime rates in making decisions concerning the shape of Federal legislation, particularly in the periodic LEAA authorizations.

Many LEAA program decisions have been based on information concerning crime and justice. The LEAA high impact crime reduction program, a $160