Are we more or less dependent on govern­ment-specifically, the federal government- today than we were 40 years ago? Has the reli­ance on the income and social support pro­grams of the federal government been a steadily increasing aspect of American life driven largely by demographic changes, or has the level of dependency changed according to the philosophy of the President in office?

Put another way, how much have federal social programs grown in areas where they might crowd out similar programs and efforts by local governments, community groups, and family networks? Have the civil society and the infrastructure of local government that sur­rounds it, in other words, yielded significant ground to a federal public sector?

These questions address a number of diffi­cult social, cultural, and economic problems that continue to challenge social scientists. Our own work elucidates and clarifies aspects of the process of mutual and reciprocal aid provided by communities yielding to federal programs, but it does not pretend to penetrate the underlying causes for this gradual but sig­nificant substitution.

The purpose of the Index of Dependency is twofold. First, the Index measures the pace at which federal government services and pro­grams are growing in areas where private or community-based services and programs exist that address the same or nearly the same needs. Second, the Index is designed to permit straightforward forecasting of its trend over the next 20 years, thus giving policy analysts the ability to estimate the degree to which changed public policy will affect the level and drift of the Index's values.

To accomplish both these goals, analysts within The Heritage Foundation's Center for Data Analysis constructed the Index in two parts: (1) a rich historical database that is employed to study the relationship between policy changes and other economic and social indicators and (2) a set of equations that ana­lysts can use to forecast future values of the Index through the year 2020.

THEORETICAL MOTIVATION

The Index uses data drawn from a carefully selected set of federally funded programs. Pro­grams were selected in terms of their propen­sity to duplicate or substitute for support given by families, local organizations, and communi­ties to people in need. Every community can point to initiatives and projects designed to assist people without adequate shelter, food, income, education, or employment. In times past, local entities provided more instances of assistance than they do today. Consequently, these local organizations played a larger role in the lives of needy people.

Over the course of the 20th century, govern­ment came to provide more and more of the services that previously had been provided by self-help and mutual aid organizations. Ade­quate housing certainly is one of the most ele­mental necessities, and mutual aid, religious, and educational organizations long had provided limited housing assistance. After World War II, the federal and state governments began to provide the bulk of low-cost housing. Today, nearly all housing assistance comes from government.

This shift from local, community-based, mutual-aid assistance to government assistance altered the relationship between the person in need and the service provider. In the past, the per­son in need depended for help on people and organizations in his or her community. The com­munity knew the person's needs and tailored assis­tance to meet those needs within the budgetary constraints of the community. Today, housing and other needs are met by government workers who frequently have no tie to the community where the needy person lives and know nothing about his or her needs beyond what they learn in the periodic reports of social workers.

In both cases, a dependent relationship exists. The first, however, is a dependent relationship with the civil society that includes expectations of the person's future ability to aid another person. The latter is a dependent relationship with a polit­ical system without any reciprocal expectations. The former is based on mutual and reciprocal aid with future aid dependent upon returning to civil viability, which, in turn, is essential to the life of civil society itself. The latter is based on unilateral aid in which the return to civil viability is not essential. Indeed, success in the latter, political system frequently is measured by the growth of the aid program.

Beyond housing, aid providers traditionally have focused on a handful of needs. Obviously, health is nearly as important as shelter, and early mutual aid societies frequently provided low-cost health care for their members. In the pre-World War II health care landscape, the indigent who were not members of these societies typically were provided health care assistance through churches and social clubs.

Virtually that entire health care infrastructure for the poor has been replaced by publicly pro­vided health care coverage, largely Medicaid and Medicare. Analysts doubtlessly are correct in not­ing that Medicaid and Medicare do a more com­prehensive job of meeting the health care needs of indigent and elderly Americans than the earlier, largely community-based systems. However, little notice has been paid to the consequences of the shifting relationship between the person receiving health care assistance and those paying for it. Today, the dependence of low-income and elderly Americans on government-provided income sup­port and health care is nearly complete. The pro­tection against policy change provided to the government health care system's patients is so strong that virtually no changes other than increases in coverage are permitted. In short, the measure of success is growth of the program.

Next to housing and health care, income assis­tance during a person's working life and during retirement clearly addresses elementary needs. Again, local, community-based charity nearly always included an income component or near-income substitutes, like free food. This element of mutual and reciprocal aid continues to have local prominence; but as with its housing and health care counterparts, government's role has become dominant since World War II. Today, Social Secu­rity provides nearly all of the income in indigent and near-indigent households. Temporary Assis­tance to Needy Families and other supplemental income insurance programs (both federal and state) address the income needs of low-income workers. Unemployment insurance payments pro­vide nearly all of the income to temporarily unem­ployed workers that once was provided by unions, friendly societies, and local charities. Indeed, income assistance is quickly becoming a govern­ment program with little if any connection to the local civil society.

COMPONENTS OF THE INDEX

The Index consists of five broad categories of programs:

• Housing assistance
• Health and welfare support
• Retirement income
• Educational subsidies at the post-secondary level, and
• Rural and agricultural services.

Federal programs and state activities supported by federal appropriations were selected to fit in each category. To be included in the initial data set, each program had to meet the standards set up by the definition of dependency: A reasonable argu­ment could be made that publicly provided goods or services could crowd out or constrain private or local government alternatives, and the immediate beneficiary had to be an individual. This standard ruled out any expenditure by the states on pro­grams that would otherwise meet the definition of dependency. However, federally funded programs, where the state acted as an intermediary, are included.

Elementary and secondary education is the principal state-based program excluded under this stipulation. Post-secondary education is the only part of govern­ment-provided education included in the Index.

Also excluded are military and federal employ­ees. Once again, national defense is viewed as a primary function of the govern­ment and thus does not promote dependency in the sense used in this research. The Index excludes non-military fed­eral employees as well on the grounds that it measures the growth of pro­grams that substitute for or crowd out similar efforts at the community level.

After this initial analysis, the data were further examined to identify the components that contrib­uted to variability. Short-term programs were elim­inated, as were relatively small programs that required little funding. The remaining expendi­tures were summed on an annual basis and divided into the five major categories listed above.

Table 1 lists the individual components of each category.^[1] The program titles in this table are those used by the Office of Management and Budget for budget function and subfunction in the budget accounting system.

Data were collected for federal fiscal years 1962 through 2002. Deflators centered on 2000 were employed to adjust for changes in the general price level, thus producing a series of real or infla­tion-adjusted values for each program listed in Table 1.

Indexes are intended to provide insight into phenomena that are either so detailed or so com­plicated that simplification through arbitrary but reasonable rules is required for obtaining anything other than a rudimentary understanding. Thus, the Consumer Price Index (CPI) of the Bureau of Labor Statistics is a series based on an arbitrarily selected "basket of goods" that the Bureau surveys periodically for changes in prices. The compo­nents of this "basket" are weighted to reflect their relative importance to overall price change, which results, for example, in energy prices being more important than clothing prices. Multiplying the weight times the price produces a weighted price for each element of the CPI, and summing up all of the weighted prices produces, roughly, the CPI score.

The Index of Dependency generally works the same way. The raw (or unweighted) value for each program (the yearly expenditures on that pro­gram) is multiplied by a weight. Summing up these weighted values produces the Index value for that year.

The Index uses the following weights:

• Housing assistance: 30 percent;
• Health and welfare support: 25 percent;
• Retirement income: 20 percent;
• Educational subsidies at the post-secondary level: 15 percent; and
• Rural and agricultural services: 10 percent.

The weights, which sum to 100 percent, are "centered" on the year 1980. This means that the weighted values for the Index components will sum to 100 for 1980, thus giving the Index a refer­ence year from which all other Index values will be evaluated.

The year 1980 was chosen because of its appar­ent significance in the history of American political philosophy. Many students of American politics believe that historians one day will view 1980 as a watershed year in U.S. history-a year that marks the beginning of the decline in left-of-center pub­lic policy and the emergence of right-of-center challenges to policies based on the belief that social systems fail without the guiding hand of government.

The Index certainly reflects such a watershed. Table 2 shows weighted values for the components and for the Index for each year between 1962 and 2002. Chart 1 plots the Index over this same time period. The drift of the scores clearly is upward over the entire period with the exception of two periods: 1980 through 1988 (Index scores of 100 and 101) and 1994 through 2000 (Index scores of 125 and 123).

These two "plateaus" in the drift of the data sug­gest that policy change may have a significant influence on the growth rate of the Index. After all, during the early 1980s, the rate of growth of some domestic programs slowed to pay for increased defense spending, and the 1990s saw significant policy changes in welfare and public housing, all of which reduced the growth rate of the Index.

Table 3 connects the Index to presidential terms. This table contains the percentage change in the Index values from the beginning of a presi­dential term to its ending. For example, the Index stood at 80.6 in 1977 when President Jimmy Carter took office. It had risen to 100 by the end of his Administration. The percentage change in the Index over that period is 24.1 percent, which is shown in the right-hand column.

It is hardly surprising that the largest jump in the Index occurred during the Administration of Lyndon Johnson. Not only did the Johnson Administration launch Medicare and other health programs, but it also vastly expanded the federal role in providing and financing low-income hous­ing. It is somewhat more surprising that the Index jumped 122 percent (from 34.7 to 77.1) under Richard Nixon and Gerald Ford. However, during these years, Republicans were funding and imple­menting substantial portions of Johnson's Great Society programs.

The two periods of more conservative public policy with respect to components of the Index stand out clearly in Table 3 and Chart 2. The slow­downs in spending increases during the Adminis­tration of Ronald Reagan and after the 1994 elections produced two periods of slightly negative change in the Index. The return of budget sur­pluses during the last years of Bill Clinton's Administration led to significant increases in spending for all of the components, particularly education and health care. Thus, the Index has resumed growing at roughly the average pace of the past 25 years. Chart 3 contains forecasted val­ues for the Index through 2020.

Forecasted Index values are based on equations employing independent variables collected on the same annual basis as the historical Index. These variables range from population figures to dummy variables that mark major legislative changes. The bulk of the data was obtained from publicly avail­able data sources, such as the Economic Report of the President, and from forecasts prepared by DRI/ WEFA, an economics consulting firm located in Lexington, Massachusetts. The dependent and independent variables were adjusted for inflation using deflators centered on the year 2000.^[2]

Using the historical data from 1962-1999, lin­ear regressions were run in each individual cate­gory. Once a model was established, each category was forecasted to the year 2020, then summed to obtain overall federal expenditures. These outlay forecasts then were converted to Index numbers using the same weights and conversion rules as employed in constructing the historically based Index.^[3]

CALCULATION OF COVERED POPULATION

The Index reflects the growth of federal govern­ment programs that arguably crowd out or substi­tute for similar initiatives at lower levels of government or within the organizations of civil society. While Index values do not depend on the number of people who receive support through these programs, that number nevertheless sheds additional light on what the Index shows.

Data on the number of people enrolled or bene­fiting from the programs listed in Table 1 between 1962 and 2001 were drawn from a variety of pub­lic sources. A significant effort was made to elimi­nate duplicate enrollments: For example, many people who receive food stamps also receive their medical services through Medicaid. Despite this effort, duplicates undoubtedly remained, and an arbitrary reduction of 5 percent in each year was imposed to account for this undetected double counting.

Chart 4 shows the annual amount of program participants from 1962 through 2001. On the eve of the Great Society programs, some 18 million people received assistance through those programs listed in Table 1 that existed at that time. Today, 17 percent of the total U.S. population-49.7 million people-receive some level of assistance through the programs covered by the Index.

Growth in income and non-financial support among program participants has accompanied the expansion of people receiving assistance. As Chart 5 shows, inflation-adjusted, per capita support (both financial and non-financial) stood at about $10,000 in 1966. By 2001, this support had grown to nearly $24,000.

Data in the Index and complementary estimates of program populations raise concerns about the ability of local governments and civil society orga­nizations to provide aid and other assistance. They raise as well a traditional republican concern about the long-term viability of political institutions when a significant portion of the population becomes dependent on government for most or all of their income.^[4]

One out of six Americans (or 17 percent) may or may not be sufficiently high to trigger this con­cern. However, this percentage grows to 24.8 per­cent when the number of federal and state employees is added to the population of Ameri­cans receiving aid through Index programs. In 1962, the sum of these two categories (Index par­ticipants and government employees) stood at 26.9 million. As Chart 6 shows, the estimate had grown to 70.6 million by the end of 2001-an increase of 162 percent since the early 1960s. This percentage growth is three times the growth rate in the U.S. population over this same period and twice the growth rate of the population aged 65 and above.

As Chart 7 shows, the annual growth rate in federal and state government employment has generally subsided since the 1960s and 1970s. However, the growth rate of state government employment has been positive for all but three years out of the past 39. Federal employment grew during the military buildup of the 1980s, and the military downsizing following the breakup of the Soviet Union and its empire led to negative change rates in federal employment throughout the 1990s.

CONCLUSION

Public policies appear to matter in the growth of the Index of Dependency. Its rapid increase in the 1960s and 1970s marked the federal government's commitment to solving local social and economic problems that previously had been the responsibil­ity of local governments, civil society organiza­tions, and families. As Chart 8 shows, the annual growth rate in the sum of government employees and the population covered by Index programs grew dramatically, even if one accounts for the mil­itary buildup of the Vietnam War in the middle years of the 1960s.

As Charts 1 and 8 show, the 1980s and 1990s generally witnessed much slower growth in the Index. Indeed, the Index would have declined over this 20-year period if the first Bush Adminis­tration had not deviated from the policies of the Reagan years and the post-1994 Republican Con­gress. However, rather than fall, the Index appears to have regained the growth rates it maintained during the Administrations of Jimmy Carter and George H. W. Bush.

While this rising Index of Dependency appears to owe its recent increases mostly to the spending opportunities provided by budget surpluses rather than to dramatic reversals in conservative public policy, several key policy debates of the next few years-for example, the debates on welfare reform, federal support for higher education, and health care reform-will likely determine the Index's rate of change for the next decade, if not well beyond.

William W. Beach is Director of the Center for Data Analysis at The Heritage Foundation.

---

^[1]Expenditure data for the Index were taken from Office of Management and Budget, Budget of the U.S. Government, Fiscal Year 2003: Historical Tables (Washington, D.C.: U.S. Government Printing Office, 2002), and Budget of the U.S. Government, Fiscal Year 2003 (Washington, D.C.: U.S. Government Printing Office, 2002).

^[2]The Center for Data Analysis at The Heritage Foundation used the Mark 11 U.S. Macro Model of WEFA, Inc., formerly Wharton Econometric Forecasting Associates. The model was developed in the late 1960s by Nobel Prize-winning econo­mist Lawrence Klein and several of his colleagues at the University of Pennsylvania's Wharton School of Business. It is widely used by Fortune 500 companies, prominent federal agencies, and economic forecasting departments. The method­ologies, assumptions, conclusions, and opinions herein are entirely the work of Heritage Foundation analysts. They have not been endorsed by, nor do they necessarily reflect the views of, the owners of the model.

^[3]Model specifications and regression results are available upon request.

^[4]For histories of this republican concern, see Bernard Bailyn, The Ideological Origins of the American Revolution (Cambridge, Mass.: Harvard University Press, 1967), and Gordon S. Wood, The Creation of the American Republic, 1776-1787 (Chapel Hill, N.C.: University of North Carolina Press, 1969).