Giving to religion: How generous are we?

There is good reason to believe that Americans are receiving an overly optimistic picture of charitable giving in the U.S. As a result, church members do not have accurate information to help assess their own giving or to evaluate broader trends. Solid information about giving is also crucial in light of current discussion about the private sector’s ability to take on social service programs that governments are relinquishing. As Arthur Hays Sulzberger, former publisher of the New York Times, said, “A man’s judgment cannot be better than the information on which he has based it.”

The main source for data on annual giving is Giving USA, a publication of the American Association of Fund-Raising Counsel, Inc. Trust for Philanthropy, which is an association of professional fund raisers. The association’s multidecade series of reports on philanthropy began in the 1950s. The Giving USA series is considered the authoritative source. It is cited under “Philanthropy” in the government’s Statistical Abstract of the United States. AAFRC data also holds a definitive place in print media ranging from the Chronicle of Philanthropy and the NonProfit Times to the Associated Press, Newsweek, Time, U.S. News & World Report, Forbes, New York Times, Chicago Tribune and Washington Post. In 1992, for example, when George Will argued that “in the greed-soggy 1980s charitable giving by individuals and corporations increased dramatically,” he was using an analysis of AAFRC data.

The problem is that AAFRC’s interpretation of the level of charitable giving in this country is almost certainly inflated. We reach this conclusion for three reasons.

•    AAFRC’s estimate of religious giving is too high, which, by extension, signals a too high estimate in overall giving.

•    By presenting certain data only in aggregate form, AAFRC’s reports do not take into account significant changes in population and individual taxes. What can be presented as a rise in giving in aggregate form may actually appear as a decline in giving when presented, for example, on a per capita basis.

•    AAFRC’s figure for individual contributions for the most recent years is a projection based on recent changes in personal income and the stock market--though AAFRC presents this as though it were actual giving data.

Before exploring the problems with the data, it’s important to outline AAFRC’s approach.

Each year AAFRC develops two independent estimates of charitable giving. One is for total charitable giving by “source”--that is, giving coming from individuals, corporations, foundations and bequests. The other estimate appears under the heading of “use,” by which they mean donations categorized by type of recipient institution, such as religion, education, health or human services. By surveying data from selected agencies that fall under that rubric, AAFRC extrapolates a total income figure for a specific “use” category. Any difference between the sum of “source” giving estimates and the sum of “use” giving estimates is assigned to a category labeled “unclassified.” Through this practice AAFRC reconciles the two overall totals.

A second decisive aspect of AAFRC’s approach is the way it treats contributions to religion. Beginning in some unstated year up through 1986, AAFRC regarded religion as what it termed a “residual” category. That meant, as the editor explained in Giving USA 1988, that “each year’s [religion] figure is the amount that remains of total giving after all other recipient estimates have been devised.” In other words, through 1986 religious giving was an estimate based on a number that remained after other giving data had been accounted for. In 1986 that residual sum was $41.7 billion.

For the years 1987 through 1995, according to the methodology used in Giving USA 1997, AAFRC updated its figure for religious giving by using denominational data to show the change in giving (by percentage) from year to year. Essentially, AAFRC used the 1986 “residual” figure for religious giving as a base point, then superimposed denominational statistics to build estimates for future years on that figure. The accuracy of the 1986 figure, of course, is crucially important. For 1996 AAFRC explained that, since denominational data were not yet published, an unspecified “relationship between personal income and giving to religion was used to estimate giving to religion.” It’s important to stress that, despite the incorporation of denominational data in these more recent years, these subsequent estimates, including AAFRC’s estimate that $66.26 billion was given to religion in 1995, still build on the earlier years when religion was treated as a “residual” category.

This brings us to our first major reason for questioning the accuracy of AAFRC’s account of charitable giving: its estimate of religious giving is too high. The accuracy of the $66.26 billion figure for 1995 can be tested by making use of another data source--one that, intriguingly, AAFRC itself uses in categories other than religion. AAFRC now keys its 1960-1972 figures for charitable giving in the areas of education and health--the second and third largest areas of charitable giving, religion being the first--to figures developed by the Commission on Private Philanthropy and Public Needs, commonly referred to as the Filer Commission after its chair, John H. Filer. Begun as a private initiative by John D. Rockefeller III with support from other national figures, including then Secretary of the Treasury William Simon, the commission’s landmark reports were published by the Department of the Treasury in the 1970s.

Of special note for the present discussion is that the Filer Commission estimated that in 1974 contributions to religion totaled $11.7 billion. As it turns out, this estimate was close to AAFRC’s 1974 estimate of religious giving--$11.84 billion.

Using the 1974 Filer Commission religion estimate of $11.7 billion as a base point, and then using available denominational data on giving, a series of annual figures for the years 1968 through 1973 and 1975 through 1995 can be calculated. (A set of 29 denominations that published data in the Yearbook of American and Canadian Churches provides the basis for calculating such an annual rate of change.) These revised numbers can be compared to the AAFRC figures.

Our Filer-based calculations yield a figure for giving to religion in 1995 of $44.47 billion. AAFRC’s 1995 estimate of $66.26 billion given to religion is 49 percent larger--a discrepancy of over $20 billion. For comparison, total giving to education, the second largest “use” category, was $17.61 billion in 1995. Figure 1 presents both the AAFRC and Filer-based religion series in aggregate current dollars.

The uncertainty of AAFRC’s figures on giving to religion points to what may be another significant problem. AAFRC suggests that more than half of all individual giving is directed to religion, and that individual giving represents 80 percent of all charitable activity. An overly high AAFRC estimate of giving to religion, therefore, raises the question of whether AAFRC’s estimate of total charitable giving is significantly above the actual mark.

The second reason we question AAFRC’s estimates on charitable giving has to do with the way it presents its numbers. When AAFRC announces charitable giving according to its “use” categories, it does so in aggregate form, without acknowledging changes in population or the amount of taxes paid. To be sure, for each “use” subcategory, such as education or health, AAFRC’s Giving USA report provides both a current dollar and an inflation-adjusted figure. However, this adjustment does not address other important factors.

Consider how Philip Meyer in The New Precision Journalism analyzes newspaper circulation for the period from 1960 to 1988. While circulation slowly increased, Meyer informs us, the population and the number of households increased faster. Therefore, while the aggregate number of newspapers sold may have increased, Meyer believes that the more telling figure reveals that the number of papers circulated per household steadily declined. The comparison to giving is obvious. While the aggregate figure for charitable giving may have risen, giving on a per capita basis may be declining.

We need, then, to analyze per-capita giving. And by considering giving as a portion of per capita disposable (after-tax) income, we can develop a picture of giving over time that takes into account both population and variation in the amount of taxes paid.

According to AAFRC’s Giving USA 1997 data, $8.42 billion was given to religion in 1968 and $66.26 billion in 1995. By taking into consideration taxes and population changes we can calculate from AAFRC data that Americans gave 1.35 percent of income to religion in 1968, and 1.25 percent in 1995--a decline of 8 percent between 1968 and 1995.

Compare these findings to our Filer-based estimate that $8 billion was given to religion in 1968 and $44.5 billion in 1995. On the basis of these figures, when population and taxes are taken into consideration, giving to religion as a percentage of income declined over these years from 1.29 percent to 0.84 percent--a decline of 35 percent from the 1968 base. Figure 2 presents per capita religion giving as a percent of disposable personal income for both the AAFRC and Filer-based numbers. The accuracy and helpfulness of estimates on charitable giving, therefore, is closely connected to how the calculations are finally reported.

Those who acknowledge a decline in religious giving often ask whether donors are redirecting money from churches to helping agencies. AAFRC reports that giving to human services was $2.31 billion in 1968 and $11.70 in 1995. If those figures are related to population figures and analyzed in terms of per capita disposable personal income, giving to human services declined from 0.37 percent in 1968 to 0.22 percent in 1995--down 41 percent from the 1968 base. That compares to an 8 percent drop in AAFRC’s own estimate of income given to religion, or the 35 percent decline in the Filer-based series, which suggests that giving as a portion of income to both churches and helping agencies has declined. (A caveat must be added here: since AAFRC does not generally provide a breakdown of how much individuals give to various “use” categories, including human services, these comparisons are only approximate.)

The third reason we question AAFRC’s estimates for charitable giving concerns the way it develops its estimate for individual giving. AAFRC’s estimate of the most recent year of individual giving is a projection--one based on changes in “total personal income, the factor given the most weight,” and “the Standard and Poors 500 Stock Index for the months of November and December.” The explanation of the methodology for Giving USA 1997 (available on request from AAFRC) indicates that the data for both 1995 and 1996 are projections. The projections, which are generally presented as though they were actual giving data, are consistently too high.

The public, then, has a difficult time evaluating a major publication such as U.S. News & World Report (December 22, 1997) when it makes the announcement--for which no source is given--that charitable activity was “up 9.5 percent in two years,” a figure that corresponds to AAFRC figures for that time period. Yet for the past six years AAFRC has routinely revised its previous year’s total charitable giving figure downward. Its market and income-based model for estimating giving is not matching reality. How much credence, then, should be placed in AAFRC’s widely reported projection about current charitable activity?

All those working in philanthropy, including AAFRC, want to provide accurate figures. Philanthropy has gained increased visibility in recent years, and we hope that the added interest can provide an impetus to continually correct and refine the information base.

To that end, we propose that AAFRC publish both the raw data and the formulae and procedures it uses to produce and update its estimates. Another important piece of information would be a table that breaks down, within each of the “use” categories, the amounts from each source category, such as individuals or corporate grants, over the time span for which AAFRC publishes data.

In the meantime, we return to some very live questions. Can the private sector absorb discontinued government programs? And how much should people give?

A guest writer in the NonProfit Times has estimated that the private sector would have to provide an additional $15.7 billion a year to make up for government cuts. If church members gave 10 percent--rather than the current 2.5 percent--of their incomes, an additional $114 billion per year could be available. If $23 billion were directed to domestic programs, there would be an additional $68 billion for international word and deed ministry available, all while maintaining current church activity.

While many people endorse the classic tithe, the data suggest that most people are not interested in giving more. On the contrary, giving has been declining as a portion of income for almost 30 years. Giving patterns will have to change dramatically for members to fulfill more of their potential to share God’s love with a hurting world. As Christians try to make intelligent decisions about these important matters, accurate information is indispensable.