Introduction

The development of long-term economic projections is not a new activity; there have been several long-term economic forecasting initiatives, including the Paley Commission effort in 1952 and the Interagency Long-Term Growth Project in the early 1960's. Recently, however, the interest in longterm growth projections has accelerated. For purposes of this review, long-term models are those which deal with developments over 5 or more years. The Office of Federal Statistical Policy and Standards with its responsibility for establishing statistical policy for all Federal Government agencies, has recognized the need for improved longterm statistical forecasting models. As a result, an Ad Hoc Interagency Committee on Long-Term Growth Projections was established in 1975 to review existing efforts in Federal agencies, to identify areas of common interest, and to examine policy options for improved coordination and integration of some of the various models.

At the outset, it is appropriate to recognize the limitations and difficulties of long-term economic forecasting. In a recent report, Data Resources, Inc. (DRI) discussed several sources of instability in the economy at present, including the disequilibrium of the international relations system, the world commodity situation, the legacy of double digit inflation, the rapid changes of relative prices, and the overall financial condition of the economy. DRI concluded that:

Under these circumstances, it is very difficult to develop serious long-range plans for government and business. Economic planning is offered as one of the solutions to our difficulties. There are longrange matters which deserve better attention from our government. But increasing frequency of shocks and the continued uncertainties make it totally inappropriate to draw up elaborate plans which assume that the future can be known. The rational strategy for businesses and governments in an environment such as this one is quite different: to develop quick responsive capabilities to new shocks as they may come along, and to devise

policies which at least partially insulate institutions and systems from the many sources of instability.' Thus, in discussing and evaluating long-term economic growth projections, one must always keep in mind the fact that, under the present economic instability, the best forecasting efforts may not be accurate enough in retrospect. It is important, however, to try to forecast the impact of current Government actions and outside events on the level of economic activity as a whole, on individual sectors and regions of the economy, and on the Federal budget in particular. Some guide to the potential effect of proposed programs derived from long-term forecasting can be an important policy tool when used in combination with other factors in comparing the impact of several possible alternative programs. Thus, the present use of long-term economic projections lies more in contributing an additional analytic dimension to the decisionmaking realm than in the area of actual knowledge of the future.

A list of several of the agencies presently involved in long-term projections illustrates the present scope of such activities within the Federal Government. The Economic Growth Branch and the Regional Economic Analysis Division of the Bureau of Economic Analysis of the Department of Commerce, the Economic Growth Division of the Bureau of Labor Statistics of the Department of Labor, the Economic Division of the Economics, Statistics, and Cooperatives Service of the Department of Agriculture, the Department of Energy (including units previously in the Federal Energy Administration and the Energy Research and Development Administration), the National Science Foundation, the Environmental Protection Agency, various Bureaus of the Department of the Interior, and the Federal Preparedness Agency of the General Services Administration all engage in long-term forecasting at some level. They have participated in the review of existing efforts presented in the next section.

Specific Project Descriptions2

BEA Growth Model

The Bureau of Economic Analysis (BEA) of the Department of Commerce is engaged in continuing development of, and projections with, a moderatesized annual growth model of the U.S. economy. The BEA Model' provides a projection of Gross National Product (GNP) and its components productivity, inflation rates, income items, and other aspects of the national economy. The BEA group maintains communications with other governmental units interested or involved in related work, especially the Bureau of Labor Statistics (BLS), in arriving at assumptions to be used for the projections.

The BEA model has also been applied within BEA to analyze the sensitivity of the economy to changes in various fiscal policy instruments and to determine capital requirements for full employment production. In addition, the model projections are used to assist other units within the Department of Commerce and other Federal agencies in their analyses of future economic conditions.

BLS Economic and Employment Projections Model

The program of economic growth studies in BLS develops 5- to 15-year economic and employment projections of the U.S. economy by industry. The projections involve a detailed study of the growth of the U.S. economy under alternative scenarios, embodying assumptions about Federal economic policy and other factors which shape the future economic environment.

Attention is given to labor force and productivity growth, capital and material requirements, and changes in technology and the patterns of demand from individuals, governments, business and foreigners. Projections of output levels as well as labor and material requirements are currently made using a 125-sector disaggregation of the U.S. economy. In addition, staff capabilities, data bases and models developed for the projection effort are regularly employed to estimate the impacts of various Federal programs, legislative proposals, and other current or anticipated developments which may affect distribution of demand, rate of economic growth, or level and distribution of employment.

? For more detailed descriptions of these models, see Joseph W. Duncan, "Long-Term Economic Growth Forecasts in the Federal Government" (1976) or Computer Simulation Methods to Aid National Growth Policy (1975).

Recent uses of the model involve projections to 1985, although projections as far as 2000 have been made.

The major use of the projections within the Department of Labor is to supply an economic and manpower framework upon which estimates of future occupational requirements are made. The Bureau of Labor Statistics regularly publishes detailed information on the outlook for employment in a large number of occupational categories.

The projections have also been used within the Labor Department and other parts of the Federal Government as a framework for assessing a number of diverse economic problems such as capital requirements, manpower utilization, and energy policy. In addition, several State and regional agencies, private research groups, and business organizations have used the projections as a "national" framework within which to develop their own, generally more disaggregated, projections. In order to make the projections as generally available as possible, a large amount of detail is published and, in addition, historical and projected data bases are made available on computer (magnetic) tape.

BEA-BLS Coordination

In the early 1960's the Interagency Growth Project (consisting of BLS, BEA and OMB, and chaired by the President's Council of Economic Advisers) guided and funded the development of a basic projection model by Dr. Lester Thurow, then at Harvard. Both BLS and BEA have enlarged and modified this basic model to reflect their separate needs for detail and focus in economic projections. They maintain close communications to ensure comparability of results from the two models in the sense that, given the same fiscal policy assumptions, the models will project the same growth rates of GNP and the same unemployment levels.

Frequently, their uses of the models differ in that BLS sets an unemployment assumption and modifies the fiscal policy assumption to achieve the assumed level of unemployment. BEA's model can be used this way, but BEA generally assumes various proposed fiscal policy packages and observes what the resulting unemployment rate would be for each.

OBERS Program

The Regional Economic staff of the Bureau of Economic Analysis has a separate projection effort in cooperation with the Economics Division (formerly Economic Research Service-ERS) of the Department of Agriculture to produce area economic projections of population, employment, personal income, and earnings for 37 industry groups. BEA produces the major economic projections, while the Economics, Statistics, and Cooperatives Service produces

projections only for the agricultural sector of the economy. This subnational projection program was begun at the request of, and with financing by, the U.S. Water Resources Council which uses the projections to assess water resources requirements and to evaluate programs. The projections involve a combination of econometric modeling and judgment.

ESCS Economic Projections Program

The agricultural portion of the OBERS projections is derived from a larger program within the Economics Division of the Economics, Statistics, and Cooperatives Service (ESCS). When the Economic Research Service (ERS) was reorganized in 1973, the National Economic Analysis Division (NEAD) was given responsibility for developing an additive, ERSwide Economic Projections Program with a quickresponse capability. NEAD has developed the core of the National-Interregional Agricultural Projections (NIRAP) system which provides OBERS data as one of its functions.

The NIRAP system is a computerized simulation of the food and fiber system, with a 10-year horizon for most projections. It can simulate alternative futures economic conditions based on scenarios. differing with respect to major uncertainties which have an impact on food and fiber and with respect to policy decisions and programs designed to alleviate specific problems. By systematic scenario development and comparative analysis of alternative future economic conditions, the range of possible adjustment paths for food and fiber can be bracketed, an early warning of potential difficulties provided, and possible solutions to potential problems and trade-offs between policy goals evaluated.

Federal Energy Administration Forecasts

Prior to creation of the Department of Energy, long-term Federal Energy Administration (FEA) projections were made through the Project Independence Evaluation System (PIES). This system generated planning estimates depicting possible states of the energy system. The model was used in two ways: (1) to help the Administrator of FEA in his policy role by analyzing the impact of various energy policies and (2) to develop a set of projections of what energy picture will be in the future. The principal result of PIES is the determination of equilibrium prices and quantities of energy by type and region at specified future time points, based on specified alternative energy policies.

Energy Research and Development Administration (ERDA) Projections

Also prior to creation of the Department of Energy, energy-related projections were published as a part of "A National Plan for Energy Research, Development and Demonstration: Creating Energy Choices for the Future." These projections were the product of a system created for the Energy Research and Development Administration (ERDA) by the Brookhaven National Laboratory (BNL), which used as its macroeconomic framework the DRI projections. The projections, which were for the years 1985 and 2000, include total energy demand, imports required, electricity used, and other factors of the national energy system, projected under a variety of scenarios.

Projections of the energy system in 1985 and 2000 based on alternative assumptions were used extensively by ERDA in developing the substance of their Plan for Energy Research, Development, and Demonstration. Implications of various alternatives, such as (1) conservation and development of greater efficiencies at end use or (2) extracting more coal and oil from current locations by developing more effective recovery technology, are examined in terms. of projected imports, demand, and other facets of the energy system. The results suggest which approaches are best for long-term and intermediate-term periods. Environmental Protection Agency

The Environmental Protection Agency (EPA) established the Strategic Environmental Assessment System (SEAS). It was a collection of interdependent models used to forecast the state of the environment which would result from alternative environmental policies and socioeconomic trends. Forecasts were presented annually through 1985. The socioeconomic trends were predicted outside the SEAS system, and the environmental policy alternatives were generated by decisionmakers in EPA.

Other Modeling Efforts

Throughout the Federal Government many agencies prepare projections about particular sectors of the economy or industries with which they are

ERDA-48, Vol. I: The Plan, GPO No. (1975) 0-579-905. More recent projections are contained in ERDA-76-;1,; A National Plan for Energy Research, Demonstration, and Development: Creating Energy Choices for the Fu ERDA-76-1, A National Plan for Energy Research, Demonstration, and Development: Creating Energy Choices for the Future. Vol. I: The Plan, GPO No. (1976) 052-01000478-6.

'It should be noted that the President's Budget for FY 1977 contains no funds or personnel for the SEAS project.

directly concerned. One example is the Federal Preparedness Agency of the General Services Administration (GSA) which develops projections of future needs for various critical materials in order to determine proper amounts to stockpile.

Another example is the Department of the Interior. Within this Department there are several long-range projection programs. These are all single-sector efforts. Some can be used as policy analysis models to see the effect on particular sectors of certain policy alternatives, but most are basically attempts at projection based on current programs. These programs include: the Minerals Availability System, projecting amounts of various minerals which will be available at certain times in the future; energy projections from the Bureau of Mines through the year 2000; regional electricity supply and demand forecasting by the regional power administrations; recreation site use forecasting by the National Park Service; and analysis of regional impact of offshore drilling by the Bureau of Land Management.

Additional Model Users: Treasury, OMB, CEA, FRB

The discussion thus far has focused on long-term economic growth models built and utilized by the Federal Government from a model builder's perspective. Many governmental groups use longterm forecasting as input to their decisions, as a basis. for policy advice to the President, or even to make projections of their own, without actually developing a large econometric model within their own agencies. Examples include the Department of the Treasury, the Federal Reserve Board (FRB), the Office of Management and Budget (OMB), and the Council of Economic Advisers (CEA). Each, for reasons that vary, finds it impractical to develop an internal longterm model, but uses the results of other modeling efforts to shape its views of the future.

Each agency has different needs for long-term projections and different ways of dealing with those needs. The CEA, for instance, is an advisory body to the President. It is asked for expert opinions on complicated economic questions, usually with a very short time to develop these opinions. Because the questions it investigates are so diverse, a model suitable for all of them would be infeasible. Instead of attempting to build an in-house model, the CEA relies on many external efforts, both public (BEA, BLS) and private (DRI, Wharton, Chase). This works well because different models are suited to answering different questions, and CEA is able to draw on the strengths of each model.

Each year the CEA, Treasury and OMB develop 5year projections of major economic variables (such as

national output, rate of inflation and unemployment rate) by extrapolating the effects of the President's proposed fiscal policy as reflected in the budget. These projections use both the BEA and private models.

The use of models by CEA and OMB offers an illustration also of how long-term models and analyses affect policy decisions. In CEA's case, a question is posed whose answer requires long-term projections. The question may be "what would be the effect of various alternative policies?" By running the proposed alternatives on an appropriate model, or on two models and combining their results, such a question can be answered fairly accurately. Some of the more special-purpose models (e.g., the energyrelated models can likewise be used to answer "what if?" questions. In addition, the projections obtained from using the models can be used as input. On the other hand, models can be used by policymakers to show what actions must be taken now to achieve a particular policy goal. Thus, for example, in the Project Independence Report, the policy goal was the achievement of energy independence. The PIES model enabled policymakers to evaluate what steps would best encourage achievement of that goal by showing how much independence could be gained, how quickly, under various possible energy policies.

The existence of several large-scale long-term modeling efforts in the Federal Government is thus useful to the policymakers of the agencies involved and also to other agencies.

Issues and Recommendations

Between mid-1975 and mid-1976 an Ad Hoc Interagency Committee on Long-Term Growth Models, chaired by the Statistical Policy Division of OMB (the predecessor of the Office of Federal Statistical Policy and Standards in the Department of Commerce), met on several occasions to exchange information on existing modeling efforts and to discuss potential areas for improved coordination. These discussions and the material presented earlier highlighted three important points:

1. A great deal of informal communication already occurs between modeling teams as a result of the need to find solutions to complex problems. Hence, a set of outputs from one model frequently becomes useful input to a second model focused on a different set of issues, so that the models are often complementary rather than redundant. Similarly, complex modeling techniques or data problems are often discussed, resulting in improvements to the various models considered.