Budget Function Classifications: Origins, Trends, and Implications for
Current Uses (Letter Report, 02/27/98, GAO/AIMD-98-67).

Pursuant to a congressional request, GAO provided information on budget
function classifications, focusing on: (1) the origins and evolution of
the current structure and recent spending trends by function; and (2)
possible challenges in using these classifications as a framework for
other governmentwide applications, such as the Federal Government
Performance Plan, required by the Government Performance and Results
Act, and the Statement Net Cost in the Consolidated Financial Statements
of the federal government.

GAO noted that: (1) while the desire to categorize federal spending
according to the purpose or mission of government can be traced back
nearly 200 years, modern budget function classifications have evolved
from a structure first used in 1948; (2) over the last 50 years, budget
functions have been generally stable, with only a few changes overall
and none in the last decade; (3) however, this period also saw
significant change in the use of budget functions from a purely
retrospective summary of federal spending, to a supplemental but
subsidiary presentation summarizing the President's budget submission,
to the basic prospective framework for the modern congressional budget
resolution; (4) over the last 20 years, federal spending has become
increasingly concentrated in just a few budget functions, with about
one-third of the functions accounting for about 90 percent of the
growth; (5) functions with the highest average annual real growth
included Medicare, Net Interest, and Health; (6) an alternative trend
analysis using a mission area structure based directly on budget
subfunction classifications shows that spending related to human
resources missions and interest payments increased from 55 percent to
nearly 70 percent of total federal spending since 1977; (7) in fact,
nearly all of the growth in federal spending since 1977 can be traced to
these two areas; (8) a more discrete analysis of federal spending based
on subfunctions affirms the prominence of interest and health-related
spending compared to other areas but also shows the rapid growth in
spending for federal law enforcement activities since 1977; (9) in
recent years, the use of budget functions for other than simple
historical summaries of federal spending--in effect, the desire to
address a variety of governmentwide needs through a convenient and
familar structure--has continued; (10) but as budget functions become
the framework for assessments of the performance and cost of government
operations, questions about the structure's appropriateness for these
emerging uses will likely increase; (11) these questions will stem from
two basic concerns: (a) how agencies report specific spending; and (b)
how this information is aggregated into various function and subfunction
categories; and (12) newly available mission and strategic goal
information arising from the Results Act and new perspectives on the
full cost of government operations arising from improvements in federal
financial reporting will present an excellent opportunity and an
appropriate environment to begin such an examination.

--------------------------- Indexing Terms -----------------------------

 REPORTNUM:  AIMD-98-67
     TITLE:  Budget Function Classifications: Origins, Trends, and 
             Implications for Current Uses
      DATE:  02/27/98
   SUBJECT:  Budget functions
             Mission budgeting
             Budget outlays
             Congressional budgets
             Agency missions
             Budget administration
             Budget authority
             Reporting requirements

             
******************************************************************
** This file contains an ASCII representation of the text of a  **
** GAO report.  Delineations within the text indicating chapter **
** titles, headings, and bullets are preserved.  Major          **
** divisions and subdivisions of the text, such as Chapters,    **
** Sections, and Appendixes, are identified by double and       **
** single lines.  The numbers on the right end of these lines   **
** indicate the position of each of the subsections in the      **
** document outline.  These numbers do NOT correspond with the  **
** page numbers of the printed product.                         **
**                                                              **
** No attempt has been made to display graphic images, although **
** figure captions are reproduced.  Tables are included, but    **
** may not resemble those in the printed version.               **
**                                                              **
** Please see the PDF (Portable Document Format) file, when     **
** available, for a complete electronic file of the printed     **
** document's contents.                                         **
**                                                              **
** A printed copy of this report may be obtained from the GAO   **
** Document Distribution Center.  For further details, please   **
** send an e-mail message to:                                   **
**                                                              **
**                                            **
**                                                              **
** with the message 'info' in the body.                         **
******************************************************************


Cover
================================================================ COVER


Report to the Chairman, Committee on the Budget, House of
Representatives

February 1998

BUDGET FUNCTION CLASSIFICATIONS -
ORIGINS, TRENDS, AND IMPLICATIONS
FOR CURRENT USES

GAO/AIMD-98-67

Budget Functions

(935223)


Abbreviations
=============================================================== ABBREV

  BEA - Budget Enforcement Act
  OMB - Office of Management and Budget

Letter
=============================================================== LETTER


B-276287

February 27, 1998

The Honorable John R.  Kasich
Chairman, Committee on the Budget
House of Representatives

Dear Mr.  Chairman: 

This letter responds to your request for information on budget
function classifications.  Budget functions have long been used as
supplementary presentations within budget and financial summaries of
the federal government.  Specifically in this letter, we discuss the
origins and evolution of the current structure and recent spending
trends by function.  In addition, as agreed with your office, we also
describe in this letter possible challenges in using these
classifications as a framework for other governmentwide applications,
such as the Federal Government Performance Plan, required by the
Government Performance and Results Act, and the Statement of Net Cost
in the Consolidated Financial Statements of the federal government. 

The budgetary information presented in this letter was developed from
an automated system used by the Office of Management and Budget (OMB)
to collect and process information for the President's annual budget
submission.  Although these data were not verified at the individual
budget account level, we did summarize and reconcile fiscal year net
outlays by subfunction to published sources.  All growth rates and
trend analyses are stated in constant dollars, using fiscal year 1996
as the base year.  Appendix I provides additional details on our
objectives, scope, and methodology. 


   RESULTS IN BRIEF
------------------------------------------------------------ Letter :1

While the desire to categorize federal spending according to the
purpose or mission of government can be traced back nearly 200 years,
modern budget function classifications have evolved from a structure
first used in 1948.  Over the last 50 years, budget functions have
been generally stable, with only a few changes overall and none in
the last decade.  However, this period also saw significant change in
the use of budget functions from a purely retrospective summary of
federal spending, to a supplemental but subsidiary presentation
summarizing the President's budget submission, to the basic
prospective framework for the modern congressional budget resolution. 

Over the last 20 years, federal spending has become increasingly
concentrated in just a few budget functions, with about one-third of
the functions accounting for about 90 percent of the growth. 
Functions with the highest average annual real growth included
Medicare, Net Interest, and Health.  An alternative trend analysis
using a mission area structure based directly on budget subfunction
classifications shows that spending related to human resources
missions and interest payments increased from 55 percent to nearly 70
percent of total federal spending since 1977 (See figure 4).  In
fact, nearly all of the growth in federal spending since 1977 can be
traced to these two areas.  Lastly, a more discrete analysis of
federal spending based on subfunctions affirms the prominence of
interest and health-related spending compared to other areas but also
shows the rapid growth in spending for federal law enforcement
activities since 1977.  Significant average annual real declines over
the last 20 years were associated with general purpose fiscal
assistance to state and local governments, certain veterans-related
activities, and the central fiscal and personnel management
activities of the federal government (See figure 5). 

In recent years, the use of budget functions for other than simple
historical summaries of federal spending--in effect, the desire to
address a variety of governmentwide needs through a convenient and
familiar structure--has continued.  But as budget functions become
the framework for assessments of the performance and cost of
government operations, questions about the structure's
appropriateness for these emerging uses will likely increase.  These
questions will stem from two basic concerns:  (1) how agencies report
specific spending and (2) how this information is aggregated into
various function and subfunction categories.  Addressing these
concerns will be challenging, but necessary, if these evolving uses
are to succeed.  Newly available mission and strategic goal
information arising from the Results Act and new perspectives on the
full cost of government operations arising from improvements in
federal financial reporting will present an excellent opportunity and
an appropriate environment to begin such an examination. 


   BACKGROUND
------------------------------------------------------------ Letter :2

Budget function classifications are intended to provide a means of
arraying budget data according to the major purpose served. 
Currently, as shown in figure 1 below, the federal budget is divided
into 20 functions to provide a coherent and comprehensive basis for
analyzing and understanding federal spending.  Of the 20 budget
functions, 17 are concerned with broad areas of national need; the
remaining three functions--Net Interest, Undistributed Offsetting
Receipts,\1 and Allowances\2 --do not address a specific national
need but are included to ensure full coverage of the federal budget. 
Subfunctions are the building blocks of functions; at present, over
70 subfunctions (see figure 1) are defined, with each intended to
describe discrete but related groupings of programs and activities. 
In essence, function classifications are merely summaries of
subfunction-based information.  To allow for an even higher level
aggregation, the function summaries are sometimes combined into six
"superfunctions:" National Defense, Human Resources, Physical
Resources, Net Interest, Other Functions, and Undistributed
Offsetting Collections. 

To facilitate preparation of the President's Budget, each budget
account is associated with a three-digit subfunction code; this
association is intended to reflect the preponderance of spending on
activities funded through that account.  All spending is totaled by
subfunction and then the subfunction totals are added to obtain
function totals.  In some cases, an account is directly assigned to a
function code if the activities funded by the account are spread
across multiple subfunctions within a single function.  In still
other cases, if the activities within the account are associated with
multiple functions, the account is associated with a generic
three-digit code ("999") to indicate that it is "multi-function." New
budget accounts are assigned a subfunction code by OMB; requests for
changes in the classification of existing accounts may be made by
each agency, subject to approval by OMB. 

Each federal activity is placed in a budget function and subfunction
that best describes its most important purpose, even though many
federal activities may serve more than one purpose.  For example,
Medicaid spending could be considered a health program or a form of
income security benefits.  To prevent double-counting and to reflect
its primary purpose of financing health care for specific
beneficiaries, the Medicaid program is classified as health care
services within the Health function.  To the extent feasible,
classifications are made without regard to agency or organizational
distinctions. 

   Figure 1:  Function and
   Subfunction Classifications

   (See figure in printed
   edition.)


--------------------
\1 Most offsetting receipts are included as deductions from outlays
in the applicable functions and subfunctions.  Some are not
distributed, however, and are aggregated into this function. 

\2 This function is used to permit the budget to reflect total
estimated budget authority and outlay requirements for future years. 


   ORIGINS AND EVOLUTION OF BUDGET
   FUNCTION CLASSIFICATIONS
------------------------------------------------------------ Letter :3

The current budget function structure, while having antecedents that
can be traced back nearly 200 years, is basically a modification of
classifications developed for the 1948 Budget.  The usefulness of
governmentwide summaries of federal spending has long been
recognized.  Throughout the 19th century, prior to the development of
federal budgeting, function presentations were devised and included
in financial summaries prepared by the Secretary of the Treasury. 
Coincident with the beginning of federal budgets covering the entire
executive branch in the early 1920s, function summaries became
associated with the President's budget submissions.  The most recent
change involving budget function classifications stems from their use
as a framework for congressional budget resolutions under the
Congressional Budget Act of 1974.  Although some changes in budget
function classifications have occurred since 1948, they have been
relatively minor; in large part, the structure has remained
remarkably stable over this period of time, with no changes in the
last decade. 

In one sense, the use of function classifications can be traced to
the earliest fiscal activities of the federal government.  The
General Appropriation Act of 1789 contained only four broad
authorizations for the expenses of the civil list, the Department of
War, discharging warrants of the Treasury, and paying pensions.  It
could be said that this act provided lump sum appropriations by
categories roughly paralleling four of today's functions:  General
Government, National Defense, Net Interest, and Veterans Benefits and
Services.  This broad form of appropriations for the executive branch
didn't last long.  In fact, only 5 years later in 1794, the Congress
had differentiated its appropriations into three separate acts, each
of which contained numerous specific (line item) appropriations.\3

The emergence of multiple line-item appropriations in the early 19th
century--often specific to the level of individual positions within
each office of the government--made it increasingly difficult to
track overall federal spending.  To address this need, the Treasury
Department developed year-end recapitulation reports.  For example,
the 1876 recapitulation was based on the following categories: 
civil, miscellaneous, and foreign intercourse; interior; military;
navy; and public debt.  The recapitulations did not appear until
after the end of the fiscal year and served only to provide summary
historical perspectives, not to provide information on financial
needs for the coming year.  Because the major purposes of federal
spending were for the most part highly correlated with the separate
agencies of government, the Treasury summaries tended to reflect the
departmental structure.  The problems of concurrence and diffusion of
missions across federal departments and agencies were not yet
significant issues. 

In 1912, President William Howard Taft's Commission on Economy and
Efficiency proposed that the President annually submit a "national
budget" and recommended that this budget contain supplemental
schedules which classified expenditures by the "amount estimated,
appropriated or expended for each function or class of work." These
schedules were intended to allow a comparison of spending across each
of the broad policy areas of the government, not just retrospectively
as in the Treasury summaries, but also prospectively by reflecting
the President's budget proposals "for the next fiscal year." The Taft
Commission's proposal also differed from earlier Treasury summaries
by introducing a multi-level classification structure--a precursor to
our modern superfunction, function, subfunction structure.  For
example, the Commission proposed three broad spending
categories--General, Public Service, and Local Government.  The
Commission's Public Service category included a Military function, as
in the Treasury recapitulations, but further subdivided this function
into "national defense by land," "national defense by sea," and
"national defense--expenditures on account of past services."
President Taft's statement to the Congress argued that these analyses
"will be of great value in considering questions of policy bearing on
the future work program of the Government."\4

Many of the recommendations of the Taft Commission were incorporated
into law as the Budget and Accounting Act of 1921.  Although this law
is the genesis for modern federal budgeting, it did not require that
a function structure be used, but only that the President provide
"supplementary information" with his annual budget submission.  The
1924 Budget--the first budget prepared after the 1921 Act--contained
a supplementary table summarizing requested and enacted
appropriations by "governmental functions." This classification was
very similar to the earlier Taft Commission proposal but also
included a "nonfunctional" category to cover items that were "of a
nonorganizational character." (e.g., interest on the debt.)

Given that the 1924 classification structure was oriented toward the
purposes of government spending, it is not surprising that it was
substantially changed as the country struggled through the dominant
crises of the next two decades:  economic depression and war.  In
fact, the 1936 Budget presented revised summaries which effectively
replaced a function orientation with one focusing on the
organizations of government (e.g., "civil departments and agencies")
and, most importantly, "recovery and relief" activities. 
Subsequently, in the 1944 Budget, all spending was recategorized as
either "war activities" or "other activities," with subordinate
detail summarized by federal department and agency.  In both cases,
the changes clearly emphasized the near singularity of purpose of
federal activities during these periods. 

The end of World War II presented an opportunity to revisit the major
purposes of the federal government, and thus the budget function
classifications.  The 1948 Budget presented a revised classification
scheme that, in effect, has formed the basis for our modern
structure.  As stated in the Budget,

     "By grouping together items which are functionally related,
     regardless of the agency that is responsible, this type of
     classification provides for the Congress and the public a useful
     summary of what the Government is doing, or expects to do, and,
     in general, focuses upon the ultimate purpose which the
     Government programs are designed to serve."\5

The revised classification included several changes intended to
promote a focus on the purposes of government.  For example, the
earlier public works function, which could be traced to the 1924
Budget, was eliminated, with spending for public works distributed
among the functions which such activities served; also, subfunctions
once again reflected the missions of government, rather than
organizations.  Another innovation of the 1948 Budget presentations
was the use of three-digit codes for all functions and subfunctions
to allow "the reader to know how each individual appropriation is
classified."

Finally, the last and arguably most significant change affecting
budget function classifications occurred in 1979, as a direct result
of the Congressional Budget Act of 1974.  This act, among other
things, created the congressional budget resolution--in effect, an
annual congressional budget plan which relates the disparate parts of
the budget to the whole and provides a means to enforce budget
targets and coordinate the budgetary actions of congressional
committees.\6 The act required that new budget authority and outlays
be shown "for each major functional category,"\7 and that changes to
budget function classifications be made "only in consultation with
the Committees on Appropriations and on the Budget of both Houses of
Congress."\8 Thus, this act defined, for the first time, a statutory
foundation for budget function classifications; most importantly, it
established budget functions as not merely residual or supplementary
schedules to aggregate historical federal spending and presidential
budget proposals, but also as the basic prospective framework for the
Congress' governmentwide budget planning. 

In keeping with the requirement of the act to present budgetary
information "in terms of a detailed structure of national needs.  . 
.[and] missions and programs of agencies,"\9 OMB proposed several
changes to function and subfunction classifications in the 1979
Budget.  In addition, a distinction was made between function and
subfunction presentations:  national needs were intended to be
described in the function classifications, with subfunctions
representing the resources devoted to agency missions. 

Although there have been periodic adjustments throughout the last two
centuries, the basic purpose of the classifications--to summarize
federal spending governmentwide--remained constant, with any
adjustments reflecting changing perspectives on the purposes of such
spending.  Since 1948, as shown in table 1 below, budget functions
have been generally stable.  Over the last 50 years, periodic changes
in the structure of budget functions occurred, but often these
changes simply reordered or renamed existing functions.  More
recently, between 1979 and 1989, two major federal activities (Social
Security and Medicare) were split into separate functions and one
(General Purpose Fiscal Assistance) was redefined as a subfunction. 
Since 1989 there have been no changes to the function
classifications. 



                                Table 1
                
                  Functions Used in Fiscal Years 1948,
                             1979, and 1998

1948                            1979                1998
------------------------------  ------------------  ------------------
National Defense                National Defense    National Defense

Veterans Benefits and Services  Veterans Benefits   Veterans Benefits
                                and Services        and Services

International Affairs and       International       International
Finance                         Affairs             Affairs

Social Welfare, Health and      Income Security     Income Security
Security

                                                    Social Security

                                Health              Health

                                                    Medicare

Housing and Community           Community and       Community and
Facilities                      Regional            Regional
                                Development         Development

Labor                           Education,          Education,
                                Training,           Training,
                                Employment, and     Employment, and
                                Social Services     Social Services

Education and General Research

                                General Science,    General Science,
                                Space, and          Space, and
                                Technology          Technology

Agriculture and Agricultural    Agriculture         Agriculture
Resources

Natural Resources not           Natural Resources   Natural Resources
primarily Agricultural          and Environment     and Environment

                                Energy              Energy

Finance, Commerce and Industry  Commerce and        Commerce and
                                Housing Credit      Housing Credit

Transportation and              Transportation      Transportation
Communication

General Government              General Government  General Government

                                General Purpose
                                Fiscal Assistance

                                Administration of   Administration of
                                Justice             Justice

Interest on the Public Debt     Interest            Net Interest

Refunds of Receipts
----------------------------------------------------------------------
The relative stability indicated by table 1 tends to mask more
frequent changes that have occurred within subfunctions.  Since 1979,
OMB has tried to use subfunctions to more discretely portray the
missions of the federal government, and thus these structures have
been subject to greater change.  For example, prior to the energy
crisis of the 1970s, all federal spending for energy programs was
included in a single subfunction with water resource programs because
most federal spending was associated with hydroelectric power
projects.  However, with the growth of other federal energy programs,
a separate energy subfunction was created.  Subsequently, this
subfunction was recast as a function, with four discrete subfunctions
(see figure 1) created to distinguish among different types of
federal missions.  While subfunction changes are driven by changes in
federal spending patterns, there is usually a lag from the point at
which spending priorities are altered to a resulting change in
subfunctions.  For example, although spending for deposit insurance
increased sharply in the late 1980s, it was not until the President's
1992 Budget that a new subfunction was added. 


--------------------
\3 For a discussion of the history of line item and lump sum budgets,
see Louis Fisher, Presidential Spending Power (Princeton, NJ: 
Princeton University Press, 1975), pp.  59-74, and Budget Object
Classification:  Origins and Recent Trends (GAO/AIMD-94-147, Sept. 
13, 1994). 

\4 Message of the President of the United States Submitting for the
Consideration of the Congress a Budget, U.S.  Senate, 62d Congress,
3d Session, Document No.  1113 (Washington, D.C., 1913), p.  24. 

\5 The Budget of the United States Government for the Fiscal Year
Ending June 30, 1948, United States Government Printing Office
(Washington, D.C., 1947), p.  1353. 

\6 Although not a public law, the budget resolution is in the form of
a concurrent resolution, which is agreed to by both Houses and thus
binding on them. 

\7 2 U.S.C.  632(a)(4).  The term "budget authority" means authority
provided by law to enter into financial obligations which will result
in immediate or future outlays. 

\8 31 U.S.C.  1104(c). 

\9 Now codified at 31 U.S.C.  1105(a)(22). 


   RECENT SPENDING TRENDS BY
   FUNCTION AND SUBFUNCTION
------------------------------------------------------------ Letter :4

Included as appendixes to this letter are several trend analyses
which show federal spending by function and subfunction for the 20
year period 1977 through 1996.  There are several ways in which
federal spending could be presented for these types of analyses; for
example, OMB publishes annual summaries showing budget authority and
net outlays.  For this letter, we developed a measure we refer to as
"adjusted net outlays." This measure is based on reported net outlays
by budget account, with one change:  we added back to net outlays
deductions for collections resulting from the business-type
transactions of the government.\10

This approach allows for a more complete analysis of changing federal
spending priorities by including all spending, whether arising from
governmental or business-type transactions with the public.  A
further discussion of this approach is included in appendix I. 

Between 1977 and 1996, average annual real growth among the functions
varied widely, ranging from -5.5 percent (General Government) to 7.4
percent (Medicare).  During the same period, adjusted net outlays for
the federal government experienced real growth of 2.4 percent, with
growth rates for the mandatory and discretionary components of the
budget at 3.3 percent and 0.8 percent, respectively.  Figure 2
displays the average annual real growth for each budget function, and
additional information by function is contained in appendix II.  As
these analyses indicate, federal spending has become increasingly
concentrated in fewer functions, with just five of the 18
functions--Medicare, Net Interest, Health, Administration of Justice,
and Social Security--accounting for about 90 percent of reported real
growth during this period. 

   Figure 2:  Average Annual Real
   Growth by Budget Function,
   Fiscal Years 1977-1996

   (See figure in printed
   edition.)

Constructing trends across budget functions can be a useful way to
observe shifting patterns in spending priorities.  However, the sheer
number of functions--17 "areas of national need" plus net
interest--and the substantial variation in absolute size across the
functions--ranging from about $350 billion to less than $15 billion
in fiscal year 1996--can present analytical difficulties. 

To address these concerns, OMB developed six "superfunctions": 
National Defense, Human Resources, Physical Resources, Other
Functions, Net Interest, and Undistributed Offsetting Receipts. 
These superfunctions are, in effect, summaries of summaries; they are
composites of function totals, which, as discussed above, are
constructed by adding subfunction totals.  Although larger and fewer
aggregations can shed light on the "macro-priorities" of the federal
government, the superfunction classifications do not really do that
because they are based directly on function definitions.  As will be
discussed in the concluding section of this letter, current budget
functions are marked by structural inconsistencies and merely adding
them together to obtain superfunction totals tends to compound the
inconsistencies. 

To allow for a more discrete analysis of the "macro-priorities" of
the federal government, while still reducing the number of
classifications needing to be examined, GAO constructed seven federal
"mission areas" based directly on subfunction totals--National
Security and International Affairs, Human Resources, Natural
Resources, Economic Affairs, General Government, Interest, and
Undistributed Offsetting Receipts.  While any aggregation is
essentially subjective, combining subfunctions directly eliminates
the arbitrariness arising from function groupings.  Each subfunction,
regardless of its function affiliation, was separately assigned to
one of the mission areas; figure 3 shows these assignments by mission
area for each of the subfunctions shown in figure 1.  The principal
differences between the GAO mission area structure and the OMB
superfunction structure occurs in the Natural Resources, Economic
Affairs and General Government mission areas.  For example within the
OMB superfunction structure, the entire budget function Commerce and
Housing Credit is placed in one superfunction, Physical Resources. 
In the GAO mission area presentation, each of the four subfunctions
within Commerce and Housing Credit is assigned to a specific mission
area:  "deposit insurance" and "other advancement of commerce" to
Economic Affairs; "postal service" to General Government; and
"mortgage credit" to Human Resources. 

   Figure 3:  Federal Mission
   Areas by Subfunction

   (See figure in printed
   edition.)

Appendix III contains trend analyses based on the GAO-defined mission
areas.  This analysis shows that average annual real growth for
interest payments (6.7 percent) and Human Resources spending (3.1
percent) dominated the period 1977 through 1996, accounting for over
95 percent of the total real growth in federal spending over the last
20 years.  Alternatively, average annual real growth in the General
Government (1.7 percent) and National Security and International
Affairs (0.5 percent) mission areas has risen, but much slower than
real growth in overall federal spending (2.4 percent); two mission
areas, Natural Resources (-0.6 percent) and Economic Affairs (-1.1
percent), experienced real declines during this period.  Figure 4
presents an alternative representation of these trends over the last
20 years.  As shares of total federal spending, interest payments
more than doubled (from over 6 percent to almost 14 percent) and
Human Resources spending increased from about 49 percent to about 56
percent.  All other mission areas experienced declines as shares of
total federal spending during this period. 

   Figure 4:  Spending Trends by
   Mission Area, Fiscal Years
   1977- 1996

   (See figure in printed
   edition.)

Appendix III also includes specific trend information on each budget
subfunction by mission area.  Figure 5 captures some of this
information by displaying the 10 subfunctions with the highest
average annual real growth and the 10 with the largest average annual
real decline, since 1977.  Although health-related and interest
spending are predictably among the fastest growing areas of federal
activity, this summary shows that federal spending for corrections
and in other law enforcement areas also experienced substantial
growth during this period.  The largest real decline over the last 20
years was in "general purpose fiscal assistance" and was the result
of terminating general revenue sharing for state and local
governments; large real declines were also associated with certain
veterans-related spending and with the central fiscal and personnel
management activities of the federal government. 

   Figure 5:  Highest Average
   Annual Real Growth or Decline
   By Subfunction, Fiscal Years
   1977-1996

   (See figure in printed
   edition.)

The preceding discussions describe the nature of spending trends by
subfunction over a 20-year period.  But subfunctions can also be used
to estimate the extent of spending change that occurs on a
year-to-year basis.\11 This type of analysis provides insight on the
relative stability in spending "according to the major purpose
served." Figure 6 displays the magnitude and frequency of annual real
change in adjusted net outlays for all subfunctions since 1977.  As
this figure indicates, most subfunction spending changes little from
year to year.  Over 40 percent of total annual real changes in
subfunction spending falls within plus or minus 5 percent; over
two-thirds of the observations fall within plus or minus 10 percent. 
The median annual real change during this period--half the
observations of real annual change in subfunction spending falling
above and half below--was about 1.5 percent. 

   Figure 6:  Distribution of
   Annual Real Changes in
   Subfunction Spending, Fiscal
   Years 1977-1996

   (See figure in printed
   edition.)


--------------------
\10 Typically, federal spending associated with the government's
business-type activities with the public is offset by revenues
collected from those activities.  Although this approach ensures that
reported net outlays reflect net governmental rather than
market-based transactions, it does understate total federal
transactions.  Reported net outlays also offsets spending arising
from interagency transfers.  To prevent double-counting both the
original and subsequent outlay, only the original outlay is included;
our adjusted net outlay totals reflect this treatment of interagency
transactions. 

\11 For a more extensive discussion of this type of analysis, see
James L.  True, Why Budgets Change a Little.  .  .Or a Lot:  The
Earthquake Budget Theory, presented at the Annual Meeting of the
Association for Budgeting and Financial Management, Washington D.C.,
October 1996. 


   NEW USES WILL POSE CHALLENGES
   FOR BUDGET FUNCTION
   CLASSIFICATIONS
------------------------------------------------------------ Letter :5

Recent federal initiatives concerned with the performance and cost of
government operations demonstrate the continued need for
governmentwide, mission-based classifications.  Although using the
well-known and available budget function classifications will clearly
benefit these initiatives initially, it will also likely surface
certain structural and procedural inconsistencies within the current
classifications.  One approach to moderate the challenges presented
by the current budget functions will be to use the results of these
new initiatives to reassess and, as needed, change the underlying
function and subfunction structures.  This approach would ensure that
the budget function classifications reflect modern statements of
federal missions and goals and, correspondingly, that an agency's
reporting of spending by function and subfunction reflects an
accurate, complete, and consistent portrayal of its missions and
activities. 

Two major and very recent developments within the federal
government--governmentwide performance plans and consolidated
financial statements--use budget function classifications to provide
a needed cross-agency, mission-based perspective. 

  -- Under the Government Performance and Results Act (the "Results
     Act"), the President is required to prepare and submit a Federal
     Government Performance Plan with the Budget submission.  To meet
     the expectation that this plan "provide a single cohesive
     picture of the annual performance goals for the fiscal year,"
     the Director of OMB was given broad discretion regarding "the
     best manner and useful form" for the plan.  Budget functions
     were selected as the principal organizing framework to summarize
     the annual performance goals of federal departments and
     agencies.  Recently, in a letter to the Director of OMB, the
     majority leadership of the Congress emphasized its expectation
     that "each budget function should have a clear mission with
     clear strategic goals" within the Federal Government Performance
     Plan. 

  -- Beginning with the prototype Consolidated Financial Statement of
     the United States Government for fiscal year 1996, a new
     presentation was added--the Statement of Net Cost.  This
     statement is intended to show the annual net cost of government
     operations and is organized around budget functions.\12

According to the Treasury, this change was made to support the
implementation of the Results Act.  Specifically, "classifying each
activity solely in the function defining its most important purpose. 
.  .
permits adding the cost of each function to obtain the total
cost."\13

These new uses will likely generate questions about the adequacy of
the current budget functions as an organizing framework.  Budget
function classifications clearly remain a convenient and well-known
mechanism for governmentwide reporting needs.  Nevertheless, current
practices regarding how information is coded within this framework,
as well as questions about both the function and subfunction
categories themselves, may focus attention on the capacity of these
structures to meet the newly evolving needs.  Specifically, (1) some
function and subfunction categories do not portray a cohesive,
mission-based grouping of federal activities and (2) agency
discretion regarding how spending is coded may not promote an
accurate, complete, or consistent portrayal of mission activities. 


--------------------
\12 "The functions in the Statement of Net Cost are the same
functions as those in the budget with one exception.  The income
security function in the budget includes federal employee retirement
and disability costs; these costs are allocated to the other major
functions in the Statement of Net Cost. 

\13 Consolidated Financial Statement, U.S.  Government, Fiscal Year
1996, U.S.  Government Printing Office (Washington, D.C., 1997), p. 
13. 


      STRUCTURAL INCONSISTENCIES
      IN BUDGET FUNCTION
      CLASSIFICATIONS
---------------------------------------------------------- Letter :5.1

A long-standing problem for any function presentation is that a
single federal activity might reasonably be associated with more than
one purpose.  For example, waste water treatment grants are directly
linked with pollution control and abatement activities but could also
be associated with area and regional development goals.  To prevent
double-counting, agencies are supposed to associate their spending
with its main or primary purpose--and budget function.  But even this
approach poses difficulties given structural inconsistencies within
current function and subfunction categories. 

For example, the 17 budget functions that are intended to indicate
specific "areas of national need" reflect a variety of organizing
principles.  While some functions comprise a set of clearly
mission-related activities (i.e., National Defense, Administration of
Justice, Health), others reflect beneficiary (e.g., Veterans Benefits
and Services), or functional themes (i.e., Transportation).  In other
cases, such as the Education, Training, Employment, and Social
Services function, an amalgamation of very different federal missions
and activities can be represented within a single budget function. 
This function includes, for example, such diverse activities as basic
knowledge and skills development (i.e, elementary through higher
education programs), vocational training and employment services, and
various social programs for children, families, and the elderly.  But
this last set of activities corresponds closely with other federal
cash and food assistance programs to needy individuals and families,
which are separately included in the Income Security function, and
with related assistance programs to specific beneficiaries such as
the elderly or veterans, which are coded to the Social Security and
Veterans Benefits and Services functions, respectively. 

Conforming agency missions and strategic goals developed under the
Results Act to current budget function classifications may not fully
capture the significance and breadth of often complex federal
entities.  The Department of Transportation offers one example. 
Virtually all departmental spending is coded to one of the modal
subfunctions (i.e., ground, air, or water transportation) within the
Transportation function.  However, the Department's recently issued
strategic plan describes strategic goals including, among others,
promoting public health and safety, enhancing economic growth and
competitiveness, and advancing the nation's vital security interests. 
Each of these goals and related agency spending could be associated
with a different, more representative mission-related budget
function. 

Although subfunctions are more homogeneous than functions, there are
comparable definitional problems.\14 For example, in the Commerce and
Housing Credit function, after excluding spending coded to the
"mortgage credit," "deposit insurance," and "postal service"
subfunctions, all remaining commerce-related spending is coded to
that function's other generic subfunction:  "other advancement of
commerce." However, this subfunction includes such dissimilar federal
activities as the Securities and Exchange Commission, the Census
Bureau, the Small Business Administration, and the National Institute
of Standards and Technology.  Perhaps more importantly, this
subfunction does not include federal efforts associated with
commercial assistance and regulation of transportation-related
activities; instead, these activities are coded to the modal
transportation subfunctions. 

An additional structural challenge arising from current budget
functions is a by-product of their most conspicuous feature--their
remarkable stability through the years.  In effect, the persistence
of the structure could become an obstacle as agencies reassesses
their missions and principal activities under the Results Act and as
new questions arise based on those reassessments.  One example of
this type of question stems from the complexities posed by fragmented
federal missions and overlapping programs.\15 For example, to compare
the performance or costs of housing programs that are spread across
several federal departments and agencies, a user would have to
consider spending that is now coded to several subfunctions in at
least four functions (i.e., Income Security, Veterans Benefits and
Services, Commerce and Housing Credit, and Community and Regional
Development).  There is no direct way using budget function
classifications to relate activities which, from a mission
perspective, are essentially comparable. 


--------------------
\14 For a detailed analysis of federal spending in fiscal year 1996
by department/subdepartment and by function/subfunction, see Budget
Issues:  Fiscal Year 1996 Agency Spending by Budget Function
(GAO/AIMD-97-95, May 13, 1997). 

\15 For a discussion of this issue, see Managing for Results:  Using
the Results Act to Address Mission Fragmentation and Program Overlap
(GAO/AIMD-97-146, Aug.  29, 1997). 


      SIMILAR ACTIVITIES MAY NOT
      BE CODED SIMILARLY
---------------------------------------------------------- Letter :5.2

Currently, agencies are expected to code each budget account
according to the preponderance of spending on activities funded
through that account.  If activities within a given account are
spread across multiple subfunctions, the account is coded to the
corresponding function; if spread across multiple functions, then a
generic multifunction code is used.  By its nature, this coding
convention tends to produce less informative results, as federal
spending is generalized to function or multifunction classifications. 
More importantly, this apparently straightforward convention does not
necessarily produce consistent results across agencies, because of
the wide range of activities that can be associated with a given
budget account and because of the discretion afforded agencies in how
they treat certain common activities. 

Even if an agency is able to associate a budget account with a
discrete subfunction, the resulting subfunction totals may not
provide a clear or complete indication of primary mission or total
cost.  For example, based on strategic goals\16 developed during its
implementation of the Results Act, the National Oceanic and
Atmospheric Administration would seem likely to associate some of its
spending with such subfunctions as "water resources," "general
science and technology," or "water transportation." However, in
fiscal year 1996, all of its spending was coded to "other natural
resources" and "other advancement of commerce," providing little if
any insight into its basic missions.  Similarly, the mission of the
National Park Service, as defined in its strategic plan, is to
"preserve unimpaired the natural and cultural resources and values of
the national park system for the enjoyment, education, and
inspiration of this and future generations." However, virtually all
of the Service's spending is coded to "recreational resources;" none
is coded to "conservation and land management."

Achieving a better understanding of the total cost of federal
operations through budget function classifications will also be
hindered because agencies sometimes code similar activities
differently.  For example, many agencies finance central support
activities through common or departmentwide budget accounts.  There
is no apparent convention for how agencies will code this type of
spending within function and subfunction classifications.  The
Department of Housing and Urban Development allocates its management
and administration accounts across four of the five subfunctions
charged by the department.\17 Conversely, the Department of Health
and Human Services places its departmental management accounts within
a single subfunction (i.e., "health care services") from among the
eight subfunctions associated with its broad mission
responsibilities.  In a related example, although all major
departments and agencies have an Office of Inspector General, these
comparable entities are coded differently.  In the departments of
Justice and Education, Inspector General spending is assigned to the
"federal law enforcement activities" subfunction; in all other cases,
Inspector General activities are coded to either one of the specific
subfunctions associated with the department's or agency's missions,
or to a general (i.e.,
"other.  .  .") subfunction. 


--------------------
\16 Advance short-term warning and forecast services; implement
seasonal to interannual climate forecasts; predict and assess decadal
to centennial climate change; promote safe navigation; build
sustainable fisheries; recover protected species; and sustain healthy
coasts. 

\17 The one subfunction used by the department for which no
allocation of central support charges was made in fiscal year 1996
accounted for only about 0.04 percent of total department
obligations. 


      ADDRESSING THE CHALLENGES OF
      BUDGET FUNCTION
      CLASSIFICATIONS
---------------------------------------------------------- Letter :5.3

The preceding discussion should not be interpreted to suggest that
budget function classifications have little or no value for the
emerging focus on governmentwide performance and cost.  On the
contrary, as a long-standing method to categorize the purpose of
government without regard to organizational arrangements, budget
functions are an obvious and available approach to apply to these
important concerns. 

Nevertheless, the structural and procedural issues described above
would seem to suggest that initial expectations should be tempered. 
For example, achieving the congressional expectation of a clear
mission and strategic goals for each budget function will likely be
very difficult; in many cases, the function structure will produce
mission and goal definitions which are (1) fragmentary, as in the
case of federal housing missions, or (2) unclear, as in the case of
functions which reflect non-mission themes. 

To gain the benefits of using a classification structure that is
well-known and currently available, while still addressing its
shortcomings, some attention could be given toward moderating the
known challenges of the current structure.  For example, focusing on
subfunction classifications, which are inherently more discrete than
function classifications, or recasting certain budget functions to
emphasize a more consistent mission focus, will help avoid potential
confusion and misinformation. 

Another perhaps more promising approach would be to use the
department and agency missions and strategic goals recently defined
under the Results Act as a baseline for comparison against current
function and subfunction classifications.  Just as the end of World
War II presented an opportunity to reassess the purposes of
government spending and led to broad changes in budget function
classifications, department and agency implementation of the Results
Act, as well as various financial management improvements\18 that are
intended to provide a better understanding of the full cost of
federal activities, will present an appropriate environment to begin
a modern reexamination.  Moreover, the convergence of these
initiatives focusing on performance and cost offers a unique
opportunity for such a reassessment by providing a complete,
consistent and fact-based foundation.  This evaluation will likely
reaffirm many of the current subfunctions, identify instances where
agency coding decisions might be changed to be clearer and more
mission-based, and disclose any significant gaps or inconsistencies
which might require new subfunction or function classifications. 
Collectively, these actions should result in function and subfunction
classifications better structured to continue their historic role as
a governmentwide reporting mechanism. 


--------------------
\18 These improvements stem from several statutes, notably the (1)
Chief Financial Officers Act of 1990, as amended by the Government
Management Reform Act of 1994, which established chief financial
officer positions across the government and required the preparation
and audit of annual financial statements, and (2) the Federal
Financial Management Improvement Act of 1996, which requires federal
financial management systems to support full disclosure of federal
financial data, including the full cost of programs and activities. 


---------------------------------------------------------- Letter :5.4

We are sending copies of this report to the Ranking Minority Member
of the House Committee on the Budget; to the Chairmen and the Ranking
Minority Members of the House Committee on Government Reform and
Oversight and its Subcommittee on Government Management, Information
and Technology, the Senate Committee on the Budget, and the Senate
Committee on Governmental Affairs; to the Director of the Office of
Management and Budget; and to other interested parties.  We will also
make copies available to others upon request. 

If you have any questions, please call me at (202) 512-9573.  Major
contributors to this report were Michael J.  Curro, Assistant
Director, and John W.  Mingus, Jr. 

Sincerely yours,

Paul L.  Posner
Director, Budget Issues


OBJECTIVES, SCOPE, AND METHODOLOGY
=========================================================== Appendix I

Our objectives were to (1) discuss the origin and evolution of budget
function and subfunction classifications, (2) describe recent federal
spending trends in the context of this framework, and (3) comment on
the implications of using this framework for modern applications. 
Information on the origins and historical development of budget
function classifications was developed primarily through a literature
search.  Comments on recent uses for these classifications were based
on published material from OMB and the Department of the Treasury. 

For appendixes II and III, we developed historical summaries based on
automated data collected by OMB.  As part of their annual budget
submissions to OMB, federal departments and agencies are required to
provide a variety of budgetary data for each budget account, the
basic building block of their submissions.  Each budget account is
required to be associated with, among other things, a specific
three-digit function or subfunction code.  OMB collects and processes
this information through the MAX budget system (formerly referred to
as the Budget Preparation System), which is used to prepare the
President's annual budget submission.  We accumulated these annual
data at the budget account level in a GAO database.  Although much of
the information provided by agencies is subject to verification
processes and edit checks by OMB, it is not audited, and we did not
independently verify the data.\1

For purposes of this analysis, we chose to focus on the 20-year
period, 1977 through 1996.  We selected this period because it was
sufficiently lengthy to show changes in federal spending patterns
while minimizing the amount of necessary data manipulation.  Because
the GAO database is a compilation of annual files, it does not
reflect changes in budget structure or concept that occur after a
specific fiscal year file has been appended.  Thus, to portray
spending trends against a consistent structure, we needed to modify
the coding of budget accounts throughout this period to ensure that
all data were coded according to the function/subfunction structure
used in the President's fiscal year 1998 budget submission.  The
process used for standardizing coding involved identifying each
account that had an "old" function code in any fiscal year prior to a
change in the function/subfunction structure and to recategorize that
account with the new code.  For example, "social security" was not a
separate subfunction prior to 1983, and its associated budget
accounts were coded to the "retirement programs" subfunction.  For
purposes of this analysis, individual "social security" budget
accounts were recoded to the current subfunction; other budget
accounts coded to "retirement programs" in earlier years were
similarly assigned to an appropriate and current classification. 

There were two exceptions to the above approach.  First, to ease
analysis, various offsetting receipt subfunctions (code 95x) were
merged into a single subfunction; certain interest subfunctions
(codes 902-908) were treated similarly.  Second, during the period
1977 through 1982, certain receipt accounts were coded to function
rather than subfunction codes; to maintain consistency with later
years, these accounts were recoded to the offsetting receipts
subfunction. 

After standardizing the function/subfunction structure, we needed to
select a unit of measure for the trend analyses.  Typically, federal
spending is described in terms of (1) budget authority--the authority
to enter into financial commitments, (2) obligations--the amount of
orders placed, contracts awarded, services rendered, or similar
financial commitments that will require payment either immediately or
in the future, or (3) outlays--the issuance of checks or disbursement
of cash to liquidate an obligation during a given fiscal year. 
Sometimes, the measure that is used is determined by the nature of
the analysis; for example, trend analyses according to budget object
class require using obligation data, because that is the only unit of
measure associated with object class presentations.  For budget
functions, although any of the above measures could be used, we chose
to concentrate on reported outlays to reflect actual "spending" in
each fiscal year, regardless of outstanding commitments or the extent
of budget authority available. 

In budget presentations in the President's Budget, departments and
agencies report outlays in both "gross" and "net" terms.  Net outlays
are gross outlays minus collections received from either federal or
nonfederal sources.  Deducting collections from federal sources
prevents double-counting on a governmentwide basis; deducting
collections from nonfederal sources ensures that reported net outlays
reflect only governmental transactions with the public.  This latter
deduction subtracts (or "offsets") from gross outlays any revenues
collected by the government as part of a business-type transaction
with the public. 

To achieve the purposes of this analysis--that is, to summarize
federal spending by "purpose" or mission--we chose to construct a
unique measure, which we refer to as "adjusted net outlays."\2 We
calculated this measure for each budget account by excluding from
reported gross outlays only collections from federal sources while
retaining collections from nonfederal sources.  Another way to
describe this measure is to say that it takes reported net outlays
and adds back collections from nonfederal sources.  To make this
calculation, we extracted the relevant budgetary data from the GAO
database and converted the data to constant 1996 dollars using the
GDP implicit deflator reported in the Fiscal Year 1998 Historical
Tables.  These constant dollar totals were then used to calculate the
share figures displayed in appendixes II and III. 

We believe that "adjusted net outlays" better ensures that the
measure of spending used in this letter fully reflects all outlays of
federal agencies, whether on not such spending arose from
governmental or business-type transactions.  The implications of this
calculation can be shown by example.  In fiscal year 1996, the Postal
Service "spent," in net outlay terms, negative $504 million--that is,
it ran a surplus of over $500 million.  In fact, gross outlays
totaled over $56.3 billion, but were offset by about $55.5 billion in
collections from nonfederal sources and about $1.4 billion in
collections from federal sources.  For this analysis, we report
adjusted net outlays for the Postal Service of about $55.0 billion
($56.3 billion minus $1.4 billion; or -$0.5 billion plus $55.5
billion). 

Adding back to reported net outlays the deductions for offsetting
collections from non-federal sources consistently increased reported
net outlays.  Table I.1 and Figure I.1 show, over the period we
examined, the relationship between net and adjusted net outlays. 



                               Table I.1
                
                 Adjusted Net Outlays and Net Outlays:
                       Fiscal Years 1977 and 1996

                                                     Billions of 1996
                                                         dollars
                                                    ------------------
                                                        1977      1996
--------------------------------------------------  --------  --------
Adjusted net outlays                                $1,122.1  $1,756.7
Net outlays                                            958.4   1,560.3
Difference (offsetting collections from nonfederal   $ 163.7   $ 196.4
 sources)
----------------------------------------------------------------------
Note:  This table includes undistributed offsetting receipts which
are excluded from analyses in appendixes II and III. 

   Figure I.1:  Net Outlays and
   Adjusted Net Outlays, Fiscal
   Years 1977-1996

   (See figure in printed
   edition.)

Although calculating adjusted net outlays affected virtually all
subfunctions, its impact was most pronounced in a rather small
subset.  Table I.2 notes that most of the difference in fiscal year
1996 between adjusted net outlays and net outlays can be traced to
just 7 of the over 70 budget subfunctions. 



                                    Table I.2
                     
                       Adjusted Net Outlays by Subfunction,
                                 Fiscal Year 1996

                              (Dollars in billions)

                                                                      Cumulative
                      Adjusted net                              percent of total
Subfunction                outlays   Net outlays    Difference        difference
--------------------  ------------  ------------  ------------  ----------------
Postal Service               $55.0         -$0.5         $55.5              28.2
Medicare                     194.3         174.2          20.0              38.5
International                 15.7          -2.2          17.8              47.5
 financial programs
Energy supply                 17.5           1.6          15.8              55.6
Mortgage credit                8.4          -4.8          13.2              62.3
Deposit insurance              2.7          -8.4          11.1              67.9
Farm income                   16.2           6.5           9.7              72.9
 stabilization
All other                  1,447.1       1,393.8          53.2              27.1
 subfunctions
================================================================================
Total                      1,756.7       1,560.3         196.4             100.0
--------------------------------------------------------------------------------
To provide further insight into the nature of any changes observed
during this period, we also extracted for each budget account the
code that is used to distinguish discretionary and mandatory spending
under the Budget Enforcement Act (BEA) of 1990.  However, because
this coding structure did not exist prior to the enactment of the
act, we determined appropriate BEA codes for all pre-1990 budget
accounts.  If a specific budget account existed both pre- and
post-1990, the post-1990 BEA code was simply added to all earlier
data.  Accounts that did not exist after 1990 were coded based on
logical matches to similar types of accounts and activities. 

For appendixes II and III, we computed real average annual growth
rates using the constant dollar totals for each budget function and
subfunction for fiscal years 1977 and 1996.  We used the "@RATE"
formula from an automated spreadsheet application to compute growth
rates.  This formula determines the periodic interest rate needed for
an investment to grow to a future value--or conversely the growth
rate displayed between two known values--over the number of
compounding periods. 

Lastly, to ensure that the process of standardizing function codes
and adding BEA codes to individual budget accounts did not affect
overall accuracy, reported net outlay totals for subfunctions and for
mandatory and discretionary spending were compared to OMB's published
sources.  After controlling for the shift of some budget accounts
between subfunctions or BEA categories, only one subfunction had
greater than a 1-percent difference between the calculated dataset
for this analysis and published data.  That subfunction, deductions
for offsetting receipts, did not significantly affect the substance
of our analysis. 

We performed our work from August through December 1997 in accordance
with generally accepted government auditing standards.  We provided a
draft of this report to OMB for technical review. 



--------------------
\1 Some recent GAO work has indicated significant problems in the
accounting information underlying certain revenue and expenditure
presentations in the budget.  See, for example, Financial Audit: 
Examination of IRS' Fiscal Year 1995 Financial Statements
(GAO/AIMD-96-101, July 11, 1996). 

\2 To create this measure, we extracted budget data from three OMB
MAX data series:  Line Type "A" (Analysis of Outlays), Line Type "P"
(Program and Financing Data), and Line Type "R" (Receipts data). 


TREND ANALYSES BY FUNCTION
========================================================== Appendix II

The 17 budget functions which reflect areas of national needs and
spending for net interest have experienced widely varying trends over
the last 20 years.  As shown in figure II.1, just 6 functions
accounted for almost 80 percent of total adjusted net outlays in
1996; those same functions accounted for about two-thirds of spending
in 1977.  As table II.1 shows, average annual growth rates varied
substantially across the functions, with most of the growth arising
from mandatory spending (i.e., spending not controlled through the
appropriations process but rather by authorizing laws which define
eligibility or set benefit or payment rules). 



                                    Table II.1
                     
                       Spending Trends and Growth Rates by
                                     Function

                           Billions of 1996       Average annual growth rate
                               dollars                    (percent)
                          ------------------  ----------------------------------
                                                          Discretionar
Budget function               1977      1996   Mandatory             y     Total
------------------------  --------  --------  ----------  ------------  ========
National defense            $235.9    $275.4        -9.1           0.9       0.8
International affairs         43.4      34.1        -1.3          -1.2      -1.3
General science, space        11.5      16.8         2.6           2.0       2.0
 and technology
Energy                        22.5      19.4         0.3          -2.3      -0.8
Natural resources and         27.5      25.0         8.7          -1.2      -0.5
 environment
Agriculture                   28.5      19.1        -2.8           1.9      -2.1
Commerce and housing          63.5      70.9         1.0          -3.4       0.6
 credit
Transportation                35.8      40.5         0.7           0.6       0.7
Community and regional        20.9      13.5        -5.4          -1.6      -2.3
 development
Education, training,          50.0      53.8         3.0          -0.4       0.4
 employment,
 and social services
Health                        43.3     124.0         7.2           1.9       5.7
Medicare                      50.5     194.3         7.5           2.6       7.4
Income security              148.0     233.0         2.0           5.7       2.4
Social security              199.2     349.7         3.1          -1.1       3.0
Veterans benefits and         46.6      39.8        -2.4           2.1      -0.8
 services
Administration of              8.6      20.0        35.9           3.9       4.6
 justice
General government            44.2      15.0       -12.7          -0.4      -5.5
Net interest                  71.4     246.0         6.7           n/a       6.7
================================================================================
Total                     $1,151.4  $1,790.5         3.3           0.8       2.4
 (excluding
 undistributed
 offsetting receipts)
--------------------------------------------------------------------------------
   Figure II.1:  Shares of Federal
   Spending by Function, Fiscal
   Years 1977-1996

   (See figure in printed
   edition.)


TREND ANALYSES BY MISSION AREA
========================================================= Appendix III

To better reflect the major spending priorities of the federal
government, GAO developed the six "mission areas" shown in table
III.1, below.  Each of these mission areas was constructed by
aggregating reported subfunction totals; table 3 in the preceding
letter, and the subsequent analyses in this appendix, define which
subfunctions are associated with which mission areas. 

As table III.1 and figure III.1 show, there have been some
significant shifts in federal spending by mission area over the last
20 years.  Human resources spending has increased from about 49
percent of total spending to almost 56 percent; in constant dollar
terms, spending has increased nearly 80 percent in this mission area. 
Even more dramatically, interest spending has increased almost 250
percent in constant dollar terms.  These two mission areas now
constitute almost 70 percent of total federal spending. 

The remainder of this appendix provides an overview of the factors
underlying the changes in each mission area. 



                                   Table III.1
                     
                      Shares of Federal Spending and Growth
                              Rates by Mission Area

                           Billions of 1996       Average annual growth rate
                               dollars                    (percent)
                          ------------------  ----------------------------------
                                                          Discretionar
Mission areas                 1977      1996   Mandatory             y     Total
------------------------  --------  --------  ----------  ------------  ========
Human resources             $563.0  $1,002.0         3.3           1.9       3.1
 [e.g., health;
 retirement and
 disability; education
 and training; income
 support]
National security and        279.3     309.5        -1.9           0.7       0.5
 international affairs
Interest                      71.4     246.0         6.7                     6.7
Economic affairs             123.5     100.3        -3.6           0.1      -1.1
 [i.e., area/rural/
 community development;
 financial institution
 oversight;
 transportation; science
 and research]
General government            64.2      88.2         1.8           1.5       1.7
 [i.e., legislative and
 executive functions;
 postal service;
 administration of
 justice]
Natural resources             50.1      44.4         1.3          -1.5      -0.6
 [i.e., energy
 conservation and
 supply; natural
 resources; environment]
================================================================================
Total                     $1,151.4  $1,790.5         3.3           0.8       2.4
 (excluding
 undistributed
 offsetting receipts)
--------------------------------------------------------------------------------
   Figure III.1:  Shares of
   Federal Spending by Mission
   Area, Fiscal Years 1977-1996

   (See figure in printed
   edition.)



   HUMAN RESOURCES
------------------------------------------------------- Appendix III:1

As noted above, the human resources mission area has been
consistently the largest mission area as a share of total federal
spending during this period and has also experienced the second
highest average annual real growth (3.1 percent).  (See table III.1.)
However, growth rates and shares of total federal spending vary
considerably across the large number of subfunctions within this
mission area and across the period of this analysis.  To help show
some of these differences, table III.2 displays four principal groups
of subfunctions within this mission area:  (1) education and
training, (2) health, (3) income support, and (4) retirement and
disability. 

Although the "social security" subfunction continues to dominate the
mission area overall (figure III.3), health-related spending--notably
"medicare" and "health care services"--has shown the most rapid real
growth during this period, now collectively representing almost 19
percent of federal spending.  (See table III.2 and figure III.2.)
Much of the increase in health-related programs--and in general for
this mission area--has arisen from mandatory spending programs,
rather than through discretionary appropriations. 

Perhaps more so than other mission areas, human resources spending
reflects significant demographic (e.g., veterans-related spending)
and economic (e.g., "unemployment compensation") changes during the
period of this analysis.  For example, the large (-5.8 percent) real
decline in the "mortgage credit" subfunction is largely driven by
interest rates associated with the selected start (unusually high
rates) and end dates (unusually low rates) for this analysis.  Also,
the decline in "training and employment" spending (-4.3 percent)
reflects the large temporary increases for public sector jobs as part
of an economic stimulus program in 1977 and 1978, coupled with the
termination of the Comprehensive Employment and Training Act after
1982; since 1983, the "training and employment" subfunction has been
virtually flat. 



                                   Table III.2
                     
                       Subfunction Shares and Growth Rates,
                           Human Resources Mission Area

                          Percent of federal      Average annual growth rate
                               spending                  (percent)\a
                          ------------------  ----------------------------------
                                                          Discretionar
Budget subfunctions           1977      1996   Mandatory             y     Total
------------------------  --------  --------  ----------  ------------  ========
[Education and training       3.98      2.20        -0.8          -0.8      -0.8
 subfunctions]
Elementary, secondary,        0.94      0.83         n/a           1.7       1.7
 and vocational
 education
Higher education              0.70      0.78        10.8           0.5       2.9
Training and employment       1.40      0.39        62.6          -5.1      -4.3
Research and general          0.18      0.12        -2.4           0.3       0.3
 education aids
Veterans education,           0.76      0.07        -9.6         -14.0      -9.6
 training,
 rehabilitation
[Health subfunctions]         9.11     18.74         7.4           2.1       6.3
Medicare                      4.38     10.85         7.5           2.6       7.4
Health care services          2.89      6.20         7.2           2.6       6.5
Hospital and medical          0.96      0.96        41.9           2.3       2.4
 care for veterans
Health research and           0.72      0.61         n/a           1.4       1.4
 training
Consumer and                  0.16      0.12        -6.4           0.8       0.8
 occupational health and
 safety
[Income support              12.01     10.14         0.6           5.2       1.4
 subfunctions]
Other income security         2.93      3.75         3.7           3.2       3.7
Food and nutrition            1.73      2.12         3.0          10.0       3.4
 assistance
Housing assistance            0.66      1.50        -6.0           7.3       6.9
Unemployment                  3.12      1.39        -2.1           0.0      -1.9
 compensation
Social services               1.04      0.83         0.4           2.3       1.1
Mortgage credit               2.28      0.47        -6.7          17.3      -5.8
Other veterans benefits       0.13      0.07        -2.2          -1.0      -1.2
 and services
Veterans housing              0.12      0.02        -8.1           n/a      -6.4
[Retirement and              23.80     24.88         2.6          -1.2       2.6
 disability
 subfunctions]
Social security              17.30     19.53         3.1          -1.1       3.0
Federal employee              3.68      3.87         2.6           2.7       2.6
 retirement and
 disability
Income security for           2.08      1.10        -1.0           n/a      -1.0
 veterans
General retirement and        0.73      0.39        -0.9          -3.1      -1.0
 disability insurance
================================================================================
Total                        48.90     55.98         3.3           1.9       3.1
 (excluding
 undistributed
 offsetting receipts)
--------------------------------------------------------------------------------
\a "n/a" in growth rate columns occurs when growth rate cannot be
measured because of sign changes in data or absence of data. 

   Figure III.2:  Human Resources
   Subfunctions as Shares of Total
   Federal Spending, Fiscal Years
   1977-1996

   (See figure in printed
   edition.)

   Figure III.3:  Shares of Human
   Resources Spending by Major
   Subfunction Groups, Fiscal
   Years 1977-1996

   (See figure in printed
   edition.)



   NATIONAL SECURITY AND
   INTERNATIONAL AFFAIRS
------------------------------------------------------- Appendix III:2

Like the human resources mission area, national security and
international affairs is dominated by a single subfunction: 
"Department of Defense - Military." (See table III.3.) Although this
subfunction has declined sharply as a share of total federal
spending--from about 20 percent to about 15 percent over the last 20
years--it has consistently maintained about 80 percent of the
spending in this mission area.  "International security assistance"
and "international financial programs" were the largest declining
subfunctions, both as a share of overall federal spending and in real
terms.  (See figure III.4.)



                                   Table III.3
                     
                       Subfunction Shares and Growth Rates,
                       National Security and International
                               Affairs Mission Area

                          Percent of federal      Average annual growth rate
                               spending                  (percent)\a
                          ------------------  ----------------------------------
                                                          Discretionar
Budget subfunction            1977      1996   Mandatory             y     Total
------------------------  --------  --------  ----------  ------------  ========
Department of Defense-       20.06     14.68       -11.1           0.7       0.7
 Military
International financial       2.15      0.88        -1.2         -11.9      -2.4
 programs
Atomic energy defense         0.39      0.65         n/a           5.1       5.1
 activities
International                 0.67      0.44        -7.2           0.3       0.1
 development and
 humanitarian assistance
International security        0.66      0.31         n/a          -1.8      -1.7
 assistance
Conduct of foreign            0.20      0.22        -6.3           2.8       2.8
 affairs
Foreign information and       0.08      0.07         4.2           1.5       1.5
 exchange activities
Defense-related               0.03      0.05         6.4           4.6       4.9
 activities
================================================================================
Total                        24.25     17.29        -1.9           0.7       0.5
 (excluding
 undistributed
 offsetting receipts)
--------------------------------------------------------------------------------
\a "n/a" in growth rate columns occurs when growth rate cannot be
measured because of sign changes in data or absence of data. 

   Figure III.4:  National
   Security and International
   Affairs Subfunctions as Shares
   of Total Federal Spending,
   Fiscal Years 1977-1996

   (See figure in printed
   edition.)



   ECONOMIC AFFAIRS
------------------------------------------------------- Appendix III:3

The overall real decline in spending for this mission area (table
III.4) reflects several significant shifts in federal priorities: 
the near ending of "general purpose fiscal assistance"--most notably
general revenue sharing and several other federal payments to state
and local governments--and a decline in regional and community
development spending.  However, overall growth rates mask several
subfunctions which experienced substantial volatility during this
period.  For example, as shown in figure III.5, the "disaster relief
and insurance" subfunction experienced several periods of rapid
increase in response to natural disasters, and "deposit insurance"
spending sharply surged in the late 1980s in response to the savings
and loan crisis. 



                                   Table III.4
                     
                       Subfunction Shares and Growth Rates,
                          Economic Affairs Mission Area

                          Percent of federal      Average annual growth rate
                               spending                  (percent)\a
                          ------------------  ----------------------------------
                                                          Discretionar
Budget subfunction            1977      1996   Mandatory             y     Total
------------------------  --------  --------  ----------  ------------  ========
Ground transportation         2.09      1.43         0.8           0.3       0.3
Farm income                   2.24      0.90        -2.8           6.0      -2.4
 stabilization
Space flight, research,       0.78      0.71      -100.0           1.9       1.9
 and supporting
 activities
Air transportation            0.58      0.57      -100.0           2.4       2.3
Area and regional             0.86      0.28        -2.8          -3.0      -3.6
 development
Community development         0.78      0.28        -7.1          -1.7      -3.0
Other advancement of          0.32      0.27        21.8          -0.9       1.4
 commerce
Water transportation          0.43      0.24         2.9          -1.1      -0.7
General science and           0.22      0.22         3.7           2.5       2.5
 basic research
Disaster relief and           0.18      0.20         3.7           2.7       2.9
 insurance
Agricultural research         0.23      0.17         1.5           0.5       0.6
 and services
Deposit insurance            -0.04      0.15         n/a           n/a       n/a
General purpose fiscal        1.96      0.11       -13.3          -9.3     -12.0
 assistance
Other labor services          0.08      0.05      -100.0           0.3       0.3
Other transportation          0.02      0.02         n/a           3.5       3.8
================================================================================
Total                        10.72      5.60        -3.6           0.1      -1.1
 (excluding
 undistributed
 offsetting receipts)
--------------------------------------------------------------------------------
\a "n/a" in growth rate columns occurs when growth rate cannot be
measured because of sign changes in data or absence of data. 

   Figure III.5:  Economic Affairs
   Subfunctions as Shares of Total
   Federal Spending, Fiscal Years
   1977-1996

   (See figure in printed
   edition.)


   GENERAL GOVERNMENT
------------------------------------------------------- Appendix III:4

Throughout the first half of the 20-year trend period of this
analysis, two subfunctions--"postal service" and "central fiscal
operations"--dominated the general government mission area.  (See
figure III.6.) Postal service spending continued throughout this
period to be a significant share of this mission area and of total
federal spending, but spending for central fiscal operations fell off
sharply.  However, this decline merely reflects a technical change in
the budgetary treatment of agency borrowing from the Federal
Financing Bank. 

Notwithstanding the dominance of postal service spending, the highest
real growth in this mission area over the last 20 years occurred in
three other subfunctions:  "federal correctional activities" (9.4
percent), "federal litigative and judicial activities" (6.2 percent),
and "federal law enforcement activities" (4.4 percent).  (See table
III.5.) Collectively, spending in these Administration of Justice
subfunctions exhibited real growth four times greater than the other
subfunctions in this mission area (4.6 percent versus 1.1 percent). 
The large real decline since 1977 in "central personnel management"
(-4.4 percent) reflects the recent downsizing of the Office of
Personnel Management. 



                                   Table III.5
                     
                       Subfunction Shares and Growth Rates,
                         General Government Mission Area

                          Percent of federal      Average annual growth rate
                               spending                  (percent)\a
                          ------------------  ----------------------------------
                                                          Discretionar
Budget subfunction            1977      1996   Mandatory             y     Total
------------------------  --------  --------  ----------  ------------  ========
Postal service                2.95      3.07         3.5         -19.6       2.6
Federal law enforcement       0.35      0.52        25.3           3.6       4.4
 activities
Central fiscal                1.53      0.49         n/a           3.0      -3.6
 operations
Federal litigative and        0.17      0.35         n/a           5.5       6.2
 judicial activities
Federal correctional          0.05      0.18         5.8           9.3       9.2
 activities
Legislative functions         0.17      0.11         n/a          -0.3      -0.1
Criminal justice              0.17      0.08         n/a          -2.3      -1.7
 assistance
General property and          0.05      0.06         6.4          -8.0       3.6
 records management
Other general government      0.09      0.06         9.3           2.9       0.1
Executive direction and       0.02      0.01         n/a           1.9       1.9
 management
Central personnel             0.02      0.01         n/a          -4.4      -4.4
 management
================================================================================
Total                         5.58      4.93         1.8           1.5       1.7
 (excluding
 undistributed
 offsetting receipts)
--------------------------------------------------------------------------------
\a "n/a" in growth rate columns occurs when growth rate cannot be
measured because of sign changes in data or absence of data. 

   Figure III.6:  General
   Government Subfunctions as
   Shares of Total Federal
   Spending, Fiscal Years
   1977-1996

   (See figure in printed
   edition.)


   NATURAL RESOURCES
------------------------------------------------------- Appendix III:5

Over the last 20 years, the natural resources mission area has
consistently been the smallest area of federal spending, with each of
the six subfunctions associated with this mission generally being
less than 1 percent of total federal outlays over this period.  (See
table III.6 and figure III.7.) Spending in this mission area was
dramatically affected by the federal response to the energy crisis in
the late 1970s, which resulted in periods of rapid growth and
subsequent decline for energy-related subfunctions.  The surge in
spending in the "emergency energy preparedness" subfunction (figure
III.7b) in 1981 was the result of oil purchases for the Strategic
Petroleum Reserve. 



                                   Table III.6
                     
                       Subfunction Shares and Growth Rates,
                          Natural Resources Mission Area

                          Percent of federal      Average annual growth rate
                               spending                  (percent)\a
                          ------------------  ----------------------------------
                                                          Discretionar
Budget subfunctions           1977      1996   Mandatory             y     Total
------------------------  --------  --------  ----------  ------------  ========
Energy supply                 1.74      0.97         0.3          -2.7      -0.7
Conservation and land         0.41      0.41        12.7           0.3       2.3
 management
Pollution control and         0.87      0.36         3.0          -2.3      -2.3
 abatement
Water resources               0.69      0.30         3.7          -2.3      -2.1
Other natural resources       0.21      0.16         4.0           0.6       1.0
Recreational resources        0.21      0.16        43.1           1.0       1.1
Energy information,           0.16      0.06         n/a          -2.6      -2.6
 policy, and regulation
Energy conservation           0.04      0.04         n/a           2.3       2.3
Emergency energy              0.03      0.01         n/a          -1.0      -1.0
 preparedness
================================================================================
Total                         4.35      2.48         1.3          -1.5      -0.6
 (excluding
 undistributed
 offsetting receipts)
--------------------------------------------------------------------------------
\a "n/a" in growth rate columns occurs when growth rate cannot be
measured because of sign changes in data or absence of data. 

   Figure III.7:  Natural
   Resources Subfunctions as
   Shares of Total Federal
   Spending, Fiscal Years
   1977-1996

   (See figure in printed
   edition.)



   INTEREST
------------------------------------------------------- Appendix III:6

This mission area, which includes interest payments on the public
debt as well as interest received by trust funds and interest paid by
the federal government, experienced the highest rate of real average
annual growth during the last 20 years, almost 7 percent.  (See table
III.7.) "Interest on the public debt" as a share of total spending
has more than doubled in constant dollar terms. 



                              Table III.7
                
                  Subfunction Shares and Growth Rates,
                         Interest Mission Area

                                      Percent of federal
                                           spending
                                      ------------------
                                                               Average
                                                                annual
                                                           growth rate
Budget subfunctions                       1977      1996     (percent)
------------------------------------  --------  --------  ------------
Interest on the public debt               8.52     19.21           6.8
Other interest                           -2.32     -5.47           7.1
======================================================================
Total (excluding undistributed            6.20     13.74           6.7
 offsetting receipts)
----------------------------------------------------------------------
Note:  All spending in this mission area is mandatory. 

   Figure III.8:  Interest
   Subfunctions as Shares of Total
   Federal Spending, Fiscal Years
   1977-1996

   (See figure in printed
   edition.)

RELATED GAO PRODUCTS

Budget Issues:  Fiscal Year 1996 Agency Spending by Budget Function
(GAO/AIMD-97-95, May 13, 1997)

Federal Fiscal Trends:  Fiscal Years 1971-1995 (GAO/AIMD-97-3,
November 1996)

Budget Function Classification:  Agency Spending by Subfunction and
Object Category, Fiscal Year 1994 (GAO/AIMD-95-116FS, May 10, 1995)

Budget Function Classification:  Agency Spending and Personnel Levels
for Fiscal Years 1994 and 1995 (GAO/AIMD-95-115FS, April 11, 1995)

*** End of document. ***