Federal, State, and Local Governments
Government Finance and Employment Classification Manual
Chapter 2 - A Brief History of Census Bureau Data Collection on Governments
Contents and Abstract:
The first attempt to collect data on the activities of state and local governments occurred during the 1840 Decennial Census; the first serious effort really did not begin until the 1850 Census. Data collection efforts expanded during the decennial censuses until the creation of a permanent census office in 1902.
Beginning in that year, the Bureau began to conduct special, if sporadic, censuses of governments up until the early 1950's, when the Bureau's authorizing legislation expanded the scope of its efforts in this field and created the quinquennial census of governments and the annual surveys conducted today.
During this time, the data classification system has evolved from one focusing on taxable wealth and debt to detailed measurements of government finance activities. Today's schema is largely identical to that put into place during the early 1950's.
2.1 Historical Background on Data Collection EffortsAlthough the first collection of data on governments occurred during the 1840 Decennial Census (numbers and types of schools and pupils), the first concerted effort began with the 1850 one. This section of the decennial census was expanded successively until the 1880 Census which resulted in a 914-page volume entitled, "Valuation, Taxation, and Public Indebtedness" (which also provided a few statistics on revenue and expenditure). Annual surveys began on a sporadic basis in 1898.
The first comprehensive compilation of governmental statistics, however, awaited the creation of a permanent Census Office in 1902. The 1902 Census of Governments led to an extensive volume on Federal, state, and local government revenue and expenditure, assessed valuations, and tax levies; public debt; and estimates of "national wealth" by state and class of property (reflecting the large dependence on property taxes). Other censuses of governments, varying in detail and scope, were conducted in 1913, 1922, 1927, 1932, and 1942.
In 1950, the Congress enacted legislation expanding the Census Bureau's responsibilities in this field (Title 13, Section 161 of the U.S. Code). It established that a census of governments be conducted every five years (those ending in "2" and "7") but ironically no funds were appropriated for the first one covering 1952. This event did not occur until 1957.
It was also during the 1950's that the Bureau began the annual finance and employment surveys generating the report series largely in existence today.
City Government Finances--Since the creation of a permanent census office in 1902, the Bureau has published city government finance data annually (except for 1914 and 1920). Coverage has varied, however. Until the mid-1950's, the survey focused on the largest municipal governments (i.e., no effort was made to estimate the financial activities of all cities). Further, before 1941 the annual series included data not only for the city corporation itself, but also overlapping jurisdictions (except county governments for cities of less than 300,000 population). Separate publications on city government finances continued until 1992-93. Since then, individual city data have been released for Census of Governments years in printed form, and annually as part of the individual unit finance data files available in public use format.
State Government Finances--Annual data collection efforts began in 1915, and have continued on a regular basis since, except for 1920, 1921, and 1933-36, when budgetary problems limited activities. In addition to the annual State Government Finances report series, the Bureau began releasing preliminary data on state tax receipts in 1939.
County Government Finances--A sample-based mail survey of county government finances was conducted for the years 1940 through 1946, providing national estimates, limited state area data, and individual county government figures for selected units. Although data were collected annually for the Government Finances reports, publication of individual county government data was suspended until the fiscal 1972-73 survey. Individual county government data were published separately from 1972-73 until 1992-93.
Public School System Finances--School system finance data have been included with the annual government finance estimates series since the inception of that series in 1952. Beginning in 1978, a separate series covering public school systems was developed, partly in response to the requests from the Department of Education. This resulted in an annual publication of state area and individual school system finance data that has continued through the present time. This survey has been expanded over the years to include additional finance transactions beyond the detail included in this Classification Manual.
Other Finance Data--The Bureau began conducting an annual sample survey on public debt in 1940 and expanded it to include revenue in 1945. Estimates of national totals by type of government were published from these surveys, which were replaced in 1952 with a broader survey canvassing data on expenditure and cash and securities as well.
Until the 1957 Census of Governments, however, this annual survey was insufficient to provide any but national estimates of financial activities. The results of the 1957 Census were used to create a more reliable sample that provided state area statistics beginning with 1958 data. This annual series has continued through 2000, although the coverage has varied slightly. For most years, estimates were developed at the state area level by type of local government. For years since 1994, however, the estimates at the state level have been aggregated for all local government, with no differentiation by type of government.
In addition to the annual surveys, the Bureau began conducting a quarterly survey on state and local government tax revenues in 1962 and on finances of major public employee retirement systems in 1968.
2.12 Employment DataCensus Bureau collection of data on public employment and payroll began in 1940 with quarterly, national summary estimates by type of government, supplemented by occasional reports for the various types of governments. Before 1946, the coverage of this survey was limited to nonschool employment of state, county, city, and township governments; data for educational employees and payrolls were estimated from data provided by the U.S. Office of Education. By 1946, the program was expanded to include special districts and school districts as well as the school employment of general purpose governments.
State area employment statistics by level of government have been issued annually since the early 1940's (except 1951) although functional detail was limited then to school vs nonschool employees. Starting with 1951 data, the Bureau collected separate data on payrolls of full- and part-time employees, thereby permitting the calculation of full-time equivalent statistics. By converting part-time employees to their full-time equivalent, these statistics provide more meaningful comparisons of employment levels among governments.
In mid-1951, the quarterly survey was modified to provide monthly figures as well, which continued through January 1955 data. After then, monthly and quarterly data on public employment have been collected by the U.S. Department of Labor with the Bureau's efforts focused on an annual sample survey and periodic census to provide national and state area data for October of each year.
Three events occurred in the late 1990's that impacted this series. No annual survey took place for 1996. Second, the reference period for the Employment survey was changed from October to March, effective with the 1997 Census of Governments. Finally, in 1999, coverage in the annual employment was expanded to develop estimates by state area by type of local government.
2.2 Evolution of Data Classification SystemThese developments in data collection on government finances and employment have also been matched by the evolution of the classification schema used. By today's standards, some of these earlier standards may seem crude at best. Also, they reflect the larger economy within which governments operate (thus, the interest in taxable property wealth in an economy that generated less cash income than today's.) Yet, the classification system that exists today was designed largely in the early part of this century and is nearly identical to that established in the early 1950's.
2.21 Finance DataAlthough data on government finances were collected as part of the decennial censuses during the previous century, the methods and definitions used produced data that are of "historical" interest, but not comparable to those collected later. For data collection efforts in this century, there are three periods within which the data collected have been comparable:
1902 to 1936 1937 to 1950 1952 et seq.1902 to 1936--While there were difference during this period in the scope and detail on wealth, debt, and taxation, the basic conceptual framework remained the same. The Bureau applied a broad definition to public debt, including such now excluded items as noninterest bearing warrants. Like today, the Bureau sought to categorize government revenues (then termed "revenue receipts") and expenditures ("cost payments"). Functional categories resembled many of those used today, except in fewer numbers. (City finances, on the other hand, provided much more functional detail on current expenditures that at present.)
1937 to 1950--Beginning in 1937 and expanded in 1941, the Bureau placed more emphasis on separately identifying intergovernmental transactions and distinguishing the different sectors of government, then designated as "general government", "enterprises", and "trust and sinking funds." Although similar in name to current categories, they were not identical in nature or application.
The concept of "enterprises", for instance, included not only government-operated liquor stores and utilities but also other types of business-like activities: toll highways, airports, and commercial activities of state universities, and the like. Furthermore, finance data for this sector were reported in commercial accounting terms (e.g., depreciation) rather than in terms consistent with the other government sectors.
Similarly, the definitions used for the "trust and sinking funds" sector differed from those used today for insurance trust statistics. Payroll taxes for unemployment compensation systems, for example, were classified as taxes rather than insurance trust revenue. The offsetting contribution to the unemployment compensation fund was categorized as a "general expenditure for contributions to trust funds."
Even the "general government" sector differed from the current one. In addition to including certain insurance trusts revenues, it covered transactions between other sectors on a "net contribution" basis. Expenditure included retirement on general debt which consisted of not only direct redemption but also any contributions to sinking funds.
One consequence of this classification schema was that the sectors of government were not additive. That is, it was impossible to determine a government's total financial activity by summing the three sectors.
1952 and later--The current classification system dates largely to that created in the early 1950's. The four sectors of government now used were created and intragovernmental transactions among them eliminated, thereby allowing their summation. The enterprise sector was eliminated except for liquor stores and the four types of utilities; the remaining commercial-like activities were subsumed into the general government sector.
The definition of intergovernmental revenue and expenditure transactions was broadened to include not just "fiscal-aid" (grants and shared taxes) but also monies for contractual services (other than as a regular customer of utilities). Another change, more in presentation than concept, was the reporting of functional expenditures in terms of total outlays rather than limited just to current operations.
One effect of these classification changes was to render all prior year data obsolete, or at least not comparable to later data. For this reason all available data for years before 1952 are based on special studies done in the 1950's to recompile pre-1952 data on a basis comparable to the current system.
2.22 Employment DataPrior to 1952, the functional detail collected on public employment was limited to distinguishing between school and nonschool employees, except for state and major city governments. Since 1952 the subclassification of employment data by function has been expanded.
As noted above, beginning in 1951 the Census Bureau started collecting payroll data separately for full- and part-time employees which allowed for the calculation of full-time equivalent statistics (see Section 5.31). Up until 1985 data, the method used to calculate these figures was based solely on payroll data. Effective with 1986 data, the annual employment survey started collecting data on the number of hours worked by part-time employees in order to use a method deemed to be a more accurate representation of full-time equivalent employment. No October 1985 full-time equivalent data are available.
See Chapter 5 for a more detailed discussion of full-time equivalent employment.