|In This Chapter||Handbook Contents | Catalog of PDF files|
For more than 35 years, the Bureau of Labor Statistics has developed medium- to long-term (10 years ahead) projections of likely employment patterns in the U.S. economy. Since the early 1970s, projections have been prepared on a 2-year cycle. The projections cover the future size and composition of the labor force, aggregate economic growth, detailed estimates of industrial production, and industrial and occupational employment. The resulting data serve the many users who need information on likely patterns of economic growth and their effects on employment. The information on future employment opportunities by occupation, for example, is used by counselors, educators, and others helping young persons choose a career, and by officials who plan education and training programs.
Projecting employment in industry and occupational detail requires an integrated projection of the total economy and its various sectors. The BLS projections are developed in a series of six steps, each of which is based on separate projection procedures and models, and various related assumptions. These six steps, or system components, deal with:
These components provide the overall analytical framework needed to develop detailed employment projections. Each component is solved sequentially, with the results of each used as input for successive components and with some results feeding back to earlier steps. Within each step, many iterations are made to ensure internal consistency as assumptions and results are reviewed and revised.
Projections of labor force participation rates for each group are developed by first estimating a trend rate of change, usually based on participation rate behavior during the prior 8-year period. Second, the rate is modified when the time-series projections for the specific group appear inconsistent with the results of cross-sectional and cohort analyses. This second step ensures consistency in the projections across the various demographic groups. Finally, the size of the anticipated labor force is calculated by multiplying the labor force participation rates by the population projections.
Aggregate economic growth
Recent projections have been based on a macroeconomic model developed by Data Resources, Inc. This model has approximately 340 behavioral equations which represent the key relationships influencing the growth and composition of the U.S. economy. The model is driven by a set of nearly 300 exogenous variables which are specified by BLS and together define a particular growth scenario for the U.S. economy.
Final demand is one way to view GDP; it is GDP distributed among final users, broadly categorized into four groups:
PCE represents demand on the part of persons and nonprofit institutions serving individuals. Rent and the imputed rental value of owner-occupied dwellings are included in this category, but the actual purchase of dwellings is classified as investment. Investment includes both fixed capital goodsthe purchase of durable equipment and structures by business and nonprofit institutionsand the value of changes in business inventories of raw materials, semifinished goods, and finished goods. Purchases by persons of owner-occupied and rental structures are also included here Foreign trade includes both exports and imports of goods and services. Government demand is defined as the goods and services purchased by all government unitslocal, State, and Federal. It includes employee compensation, but does not include transfer payments, interest payments, grants, or subsidies, all of which are accounted for under personal consumption expenditures or other categories of demand.
Final demand, along with intermediate flows of goods and services among industries, determines total output at the commodity and industry level of detail. Industry output, in turn, is the key determinant of employment requirements. Projections of demand are therefore a key element in the system since variations in the structure of the demand for goods and services, combined with changes in the means of producing these goods and services, results in changing patterns of employment demand in future time periods.
To project final demand, the same kinds of judgments and assumptions are made as those that enter into the macroeconomic model. For example, demand for residential construction at the total level depends heavily on demographic and income forecasts. Breaking total residential construction down into the components of single-family, multi-family, and mobile homes depends on the same determining variables. Judgments are also made with regard to the effect of technological developmentssuch as computers and robotson the mix of investment goods as well as on purchases by other components of final demand.
The initial projections of the various categories of final demand generated by the macroeconomic model provide a starting point for the analysts, who review all aspects of demand to ensure that the models remain balanced and consistent throughout the development of a new set of projections. Although the four basic types of final demand, personal consumption expenditures, foreign trade, investment, and government spending, are subject to different procedures, they have, for the most part, two basic steps in common. First, the detail available from the macroeconomic model is further disaggregated. For example, the 18 personal spending categories are expanded to approximately 80 more detailed product categories. The sum2 (or sums) of the more finely detailed estimates is then compared with the controls from the macroeconomic model and one or both is adjusted. There is considerable flexibility in terms of how to reconcile the competing estimates and the actual procedure used varies. Second, for each of the detailed categories of final demand, a projected distribution by commodity (bridge table) is estimated and used to allocate spending to commodities.3 Final demand is always expressed on a product, or commodity, basis. The translation from commodity demand to industry output takes place at a later stage of the projections. The basic procedure can be summarized in the following relationship:
e = vector of final demand by commodity sector
c = vector of final demand by product type
G = bridge table in which each column contains the allocation of a product type to commodities in percent terms.
The projected bridge tables (G) reflect such factors as expected changes in technology, consumer tastes or buying patterns and the industrial composition of exports. They allow the analyst to provide for shifts in the commodity makeup of a given demand category. The operative principle in this procedure is to begin with the most detailed demand estimates possible. Generally, more detailed demand categories are composed of a smaller number of commodities making the resultant bridge coefficients more stable. Having the data at this level of detail also allows more precise adjustment for changes in technology, tastes and other structural factors. The bridge table also serves a second function in converting goods expenditures from purchasers to producers value. This entails separating the producer or plant value of the commodity from the distribution and transportation activities (margins) needed to deliver it to the final consumer. Thus, the value of each category of final expenditure is ultimately allocated to one or more commodities and, where appropriate, to the trade and transportation sectors.
The demand bridge table provides the crucial link between the functional view of the economy embedded in the macroeconomic model and the sectoral view needed to develop industrial and occupational employment projections. However, before the analyst can make use of the detailed bridge tables, the expenditure data from the macroeconomic model must be disaggregated to match the product detail of the bridge tables. This is done differently for each of the major demand categories. In addition, the specific procedures used may vary from one projection study to the next as research leads to improved data and models.
The BLS input-output model consists of two basic matrices for each year, a "use" and a "make" table (expressed in coefficient form). The "use" table, the principal one, shows the purchase of commodities by each industry as inputs into its production process. In coefficient form each column of this table shows the pattern of commodity inputs per dollar of industry output. Projecting this table must take into account the changes in the input pattern or the way in which goods or services are produced by each industry. In general, two types of changes in these input patterns are made in developing a future input-output table: (a) those made to the inputs of a specific industry (as, for example, the changes in inputs in the publishing industry); and, (b) those made to the inputs of a specific commodity in all or most industries (as for example increased use of business services across a wide spectrum of industries). The "make" table shows the commodity output of each industry. It allocates commodity output to the industry to which it is primary and to all other industries where the commodity is produced as a secondary product. In coefficient terms this table shows the industrial distribution of production for each commodity. Unlike the "use" table the "make' table is generally held constant or changed very little over the projection period.
Once projected values of the "use" and "make" relationships are available the projection of commodity demand developed in preceding steps is converted into a projection of domestic industry output using the following relationship:
g = D(I - BD)-1e
g = vector of domestic industry output by sector
B = "use" table in coefficient form
D = "make" table in coefficient form
I = identity matrix
e = vector of final demand by commodity sector
This particular formulation assumes that industries produce their primary and secondary products using the same technology, the industry technology assumption. This means that the technology of industries as expressed in the "use" table is independent of the commodity distribution of their output.
The demand for wage and salary hours is projected using an estimated equation derived from the first order conditions of a constant elasticity of substitution production function modified to include a time variable. This equation relates an industry's labor demand to its output, its wage rate relative to its output price, and a time trend. Annual average weekly wage and salary hours per job are then estimated as a function of time and the unemployment rate. The projection of average hours is then used to convert the projection of wage and salary hours into jobs.
The number of self-employed and unpaid family workers (SEUFW) is derived by first extrapolating the logic of the ratio of the group to the total for each industry as a function of time and the unemployment rate. The extrapolated ratio is then used to derive the level of self-employed and unpaid family workers from the number of wage and salary jobs by first calculating the total number of jobs and then subtracting the number of wage and salary jobs from the total. The hours for self-employed and unpaid family workers are then calculated by applying their estimated annual average weekly hours to their levels. Finally, total hours for each industry is derived by summing wage and salary and self-employed and unpaid family worker hours.
The results produced by these procedures together with industry output provide a measure of labor productivity. Implied rates of change in productivity are examined closely for consistency with historical trends. At the same time attempts are made to identify industries which may be expected to deviate from past behavior because of changes in technology, demand, or other factors. Where appropriate, changes to the initial employment estimates are made either by modifying the employment demand relationships themselves or by modifying results from earlier steps of the projections process.
Because staffing patterns of industries may change over time, the projection method must account for such shifts. This is done in a series of steps. First, historical data are reviewed to identify trends. Factors underlying these trends are then identified through analytical studies of specific industries and occupations, technological change, and a wide variety of other economic data. Finally, judgments are made as to how the pattern will change in the future. Factors underlying this change are numerous, including technological developments affecting production and products, innovations in the ways business is conducted, modifications of organizational patterns, responses to government policies, and decisions to add new products and services or stop offering old ones.
Some expected trends may not be evident in the historical data. For example, an analysis of the past would not point toward the future impact of robots on staffing because this technology has not been used much in most industries. However, robots are expected to have a significant impact on some occupations, especially in the automobile and similar assembly industries. Information of this nature is identified in studies conducted by the BLS Office of Productivity and Technology as well as other research-oriented organizations.
The change projected for a specific occupation may be small, moderate, or significant; the precise percentage reflects the judgment of the staff members based on the analyses described above that relate to that occupation. In general, changes in coefficients averaging about 10 percent are considered to be small; changes of 20 percent are moderate; and changes of 30 percent or more are considered to be significant. Documentation released with the projections provides detail on the assumptions developed for each of the occupations for which changes to the base-year coefficients are made.
Once projected staffing patterns are available they are used to allocate each industry's projected employment to detailed occupations. These estimates can then be summed across industries to yield total employment for each detailed occupation as follows:
o = vector of wage and salary employment by occupation
l = vector of wage and salary employment by industry
S = staffing pattern matrix in which each column contains the allocation of industry employment to occupations in percent terms.
The estimates described above relate only to wage and salary employees. Other classes of workers, primarily the self-employed, are analyzed separately. They are then combined with wage and salary workers to produce a forecast of total occupational demand for the United States.
Bureau of Labor Statistics
Last modified: April 28, 1999
|2000-2001 Handbook Contents...||UMSL Govt. Docs...||UMSL Libraries...||UMSL Home...|