Back >

Response to The Boston Foundation Report

Re: The "Geography and Generosity" Report from Boston College and The Boston Foundation

Date: 29 December, 2005

---

I. The Report Itself: 

We welcome, and have worked to stimulate, civil discourse on philanthropy and charitable giving. We believe there are better ways to do it than the one chosen in this case.  We feel that we have little choice but to respond, given the scale and intensity of the public attack against our work. 

The Report begins ("Executive Summary", first paragraph) with a remarkable assertion: "The resolution of this issue [i.e., of disagreements about "the generosity of Massachusetts residents"] holds profound significance for the overall economy of the Commonwealth, and especially for the state's vital nonprofit sector."  There is no evidence (and none is proposed) to support this expansive claim. 

The Report has three main parts: 1) a critique of the Generosity Index (GI) and its use of IRS data, 2) an extensive and detailed development of what is called "A New Way to Measure Generosity", and 3) what is called "A New Approach to Promoting Philanthropy". 

1) The critique is expansively beside the point, since it focuses not on the GI, but on media reporting of the GI (the Report shows no awareness of the distinction) and on straight misunderstandings of the GI some of which are perhaps colored by media interpretations.  The cornerstone and fundamental flaw of the Report vis-a-vis the GI is its allegation that the GI is a "measure to compare the generosity of the residents of different states" suggesting a competition among them for first place (p.6).   This has been frequently asserted by the media, but denied by us.   

Given our denials of their basic premise (not to mention that the Commonwealth's "overall economy" and our "vital nonprofit sector" supposedly hang in the balance), one might expect from scholars careful documentation that this is indeed the GI's purpose.  But again, no evidence is proposed (and none exists).   

Instead, the Report goes on at length to show why the GI cannot possibly measure or compare people's generosity, how it is arithmetically impossible for Massachusetts or any high-income state to "attain rank 1", and how IRS data on Itemized Charitable Deductions (ICDs) do not represent all taxpayers or all donors, etc.  We entirely agree—all this is obvious; harder to understand is why it never occurred to the authors, as the case for the absurdity of their fundamental premise mounted, that perhaps they should validate it. 

2) "A New Way to Measure Generosity" offers a good example of why we have never tried to do it—because available data is inadequate to the task, and so must be supplemented by "estimates" which can lead anywhere (such as inflating the generosity of this or that State), with few or no empirical constraints, even on major points: 

"The estimate of the average contribution made by each non-itemizing household in each state was multiplied by the estimate of the number of non-itemizing households in each state to develop an estimate of the aggregate charitable contributions made by all non-itemizing households within the state."  (p. 31—Of course multiplying estimates multiplies any errors in each one.)   

The Report also adds errors, piling estimates on estimates; as one proceeds from step to step in this excursion, earlier estimates are taken as data, on which further estimates are then piled, again on crucial points: 

"By estimating the total income and total charitable contributions for each state (that's 102 estimates—the extra two are national total estimates), it is possible to then calculate the share of [estimated—gm] total income in each state and the share of [estimated—gm] total contributions made by the residents in each state." (p.17—another 100, for a total 202 estimates in this operation alone). 

Estimates often include assumptions, which seem arbitrary: 

"This analysis calculates the proportion of estimated property taxes paid by individuals [as distinct from institutions—gm] as a percentage of total property tax revenues per state, and used [sic] this fraction applied to total sales tax revenues as an estimate of those paid by individuals." (p. 20)   

State costs of living (the authors borrow this one) are inferred from data for urban and metropolitan areas—because "costs of living" refer to distinguishable economies, which do not conform to state boundaries.   

By the time all this stuff is piled up, the empirical value (i.e., truth) of the result is impossible to know or to demonstrate, and of course fine distinctions within it (such as Massachusetts ranking 11th in generosity) are close to meaningless. Its epistemological status is: compounded estimates.  Surprisingly, important parts of philanthropic generosity (e.g., volunteering) are also ignored by this "new way" of measuring.  To repeat: this illustrates why we have always declined to claim we "measure" "generosity"—because it can't validly be done.  Identifying one indicator of generosity, such as the IRS data, is a long way from measuring it.  

3) Another false allegation of the Report, based on media misunderstanding of the GI and also without any documentation tying it to the Catalogue or the GI itself, is that the GI constitutes a "scolding" approach to promoting philanthropy and increasing charitable giving.   No one who reads the Catalogue can think we scold, but the Report then claims to offer a "New Approach", which it calls the "inclination model".  That this is "new" will come as a surprise to fundraisers, for whom it has been routine practice for decades, perhaps forever.  As for the Catalogue, one of our hallmarks has been our effort to replace inadvertently negative vocabulary, conceptualization, and rhetoric of philanthropy with positive alternatives more attractive to the younger "new and emerging donors".  We are glad the authors and sponsor of this Report have become interested in promoting philanthropy, but see no need for them to deprecate existing programs, with which they show no familiarity. 
 

II. To Be Constructive: the Generosity Index and the Report Operate in Different Spheres: 

First, let us be clear: the concept or the practice of attempting to "measure" comparative philanthropic generosity among citizens of states is simply not a priority for the Catalogue for Philanthropy, which is why it is not and never has been the focus of the Generosity Index. 

Second, the reason it is not a priority for us is that we approach these issues not from an academic or theoretical standpoint, but as practitioners.  We are interested in improving and increasing charitable giving through donor education.  Our budget is limited, so we focus on potential major donors as the most cost-effective subject of our program investments; they are the smallest group with the greatest potential for growth in giving.  All fundraisers, operating on similarly limited budgets, know that most dollars (by far) come from small numbers of large donations, rather than large numbers of small donations.  Concern about whole populations, and the "science" of measuring (actually, "estimating") their generosity, is more likely to be of interest to academics or journalists, than fundraisers. 

Third, therefore Itemized Charitable Deductions (ICDs) are highly significant data for us, and not for the Report.  Though ICDs represent a minority of donors (25-30%), they represent a substantial majority of dollars—generally about 80% (in 2003 82.2% using the Giving USA estimate of total personal giving).  The Report, by contrast, focuses on the 70-75% of non-itemizers, who in 2003 contributed only 17.8% of the dollars.  We and the Report have characteristically different interests. 

For our purposes, ICDs as a data set are powerful indicators of giving precisely because they come from the bellwether donors in the top income groups and the top of each income group, who both itemize, and give, the most.  Nationwide in 2003, 91% of the top income group (>$200,000) and 86% in the second group ($100,000-$199,999) itemized charitable deductions (in Massachusetts, 95% and 90%, respectively).  Rates of itemization decline with income—e.g. to 72% in the third highest income group ($75,000-$99,999); and 30% for all taxpayers. ICDs are especially "telling" with respect to the most potent donors. 

For us, the problem with the IRS numbers is on the income, not the giving, side (for the Report, it is the opposite).  The IRS does not separate out the income of itemizers from that of all taxpayers.  To get a better sense of who the itemizers are and how their giving relates to their income, the Catalogue for Philanthropy and the National Center for Charitable Statistics (NCCS) in 1999 together purchased (for $70,000) from the IRS a unique data set for the year 1997, which summarized down to the zip-code (not just the state) level, and separated-out the income of itemizers of charitable deductions.  The Catalogue analyzed these numbers for every municipality (over 350) and zip-code (over 600) in Massachusetts, and found that on average, itemizers have twice the income of all taxpayers and three times the income of non-itemizers, and that even within the same income groups significant disparities obtained.  In other words, it is the more affluent taxpayers in general and in every income group who are both itemizing charitable deductions, and providing the greatest share of dollars (n.b., not donors) in philanthropy. 

In sum, the key difference between the GI and the Report is that we are focused more on dollars, for which the ICDs are revealing; they are focused on donors, and therefore getting beyond ICD data.  It is not appropriate, nor is there any need, for them to criticize us for not being them.      

Similarly, the Report's exertions to raise Massachusetts' statistical generosity by raising giving numbers through including non-itemized charitable donations, and to lower available-income numbers by invoking costs of living and tax burdens, are more appropriate to their interest in all donors, than to our interest in major dollars.  The average income in the top group is over $500,000, and at that level millions of dollars of investment assets also support each taxpayer's lifestyle.  This is why NewTithing research shows that the top income groups could multiply their giving without compromising their lifestyles (incidentally, NewTithing's research directly contradicts the Report's authors' contention [p.40], which they say is based on research for this Report though we don't see how, that as income ascends, so also does percentage of giving to income).  As income rises, costs of living and tax burdens fall away as impediments to charitable giving; the opposite is true as the income scale descends; so costs of living, and state and local tax burdens, are more important to the Report's interest in measuring generosity of whole state populations than to the GI's focus on the top income groups and on the upper ends of each lower income group. 

The Report and the GI therefore have different interests, which are expressed in different uses of data.  There is no need for there to be any controversy between us, and each will be more productive if we work together as colleagues. 
 

III. Do Public Discussions of Charitable Giving Increase Giving?       

The authors of the Report, and The Boston Foundation, have said publicly that we have no evidence that media discussion helps giving; that is false and unfounded—we have been talking about this evidence for six years, since it first came out in independent IRS data. 

For the period 1991-6, Massachusetts had a lock on 50th place in the GI.   

In 1997 the GI was introduced, and evoked many loud objections, fully developed in the media, both to the facts and their implications.  What happened to giving?  Massachusetts suddenly shot up to 47th place on the GI—a new record, and quite commendable movement by such a large state; apparently donors responded as grown-ups who, when given factual feedback about their behavior, change their behavior.   

In 1998 there was no media coverage ("we did that last year"), and what happened to giving?  MA dropped back into 50th place on the GI.   

In 1999 we hired a publicist to get the story out about the significant 1997 increase in MA charitable giving; he did, and what happened to giving?  MA returned to its record high, 47th place.   

In 2000, with continued strong media coverage we surged even higher, to 44th place.   

In 2001 the 4-year rise was interrupted by economic downturn and September 11th, from which we are still recovering.   

From 1997 to 2000, then, annual giving in MA doubled—from $2-$4 billion—by far the largest gain in the nation (income everywhere rose 39%; giving nationwide rose 62%, in MA 97%) and in Massachusetts' history.    

Comparing the two periods—1991-6 without the GI, and 1997-2000 with the GI, and in particular the coincidence of the pace in the second period with that of GI publicity, it is fair to say that: 1) there was no evidence that the GI reduced giving; 2) there are two indications—increase and pace of increase—that the GI may have influenced MA giving; and 3) that the influence was positive.  Such remarkable progress in the latter period also shows that Massachusetts could and did readily increase its charitable giving—in fact, the gains (or at least some of them) may have been permanent; they were not all lost in the downturn and are now well on their way back to our record high in 2000.  Finally, the gains were entirely the work of the top income group, whose share of the total ICD in those four years rose from 51% to 74%—proving that they have substantial capacity for increasing their giving, and are an appropriate audience for the Catalogue and the GI. 

We know of no other Massachusetts influence on Massachusetts giving that helps to explain the difference between what was happening to giving here, and in the rest of the nation.  We do not claim to have been the sole cause of the difference, and we would be very interested in any other explanations.  (Income, incidentally, did not rise here more than elsewhere; nor did income decline here in 1998; so our distinctive giving history is not explained by reference to income or to a distinctive income history.) 

There is no evidence of any negative effect of the GI or its media discussions on giving; nor did the Catalogue "scold" anyone—we cheered for the donors. All the evidence we know of suggests the opposite.  
 
 
 

IV. Further Values of the Generosity Index Not Mentioned in the Report; 

The Catalogue created the "Generosity Index" in 1997, as a report to the public of the best available (IRS) data on both income and charitable deductions.  We wanted to include both because generosity is not just how much one gives, but how much one gives in relation to how much one has; previous discussions mentioned only giving, mainly because income and net-worth data on donors are not readily available.  We decided to go with the best data available, being careful to point out both its limitations and its strengths.

 

In fact the Index was invented to illustrate concisely a simple fact at that point in time: that Massachusetts and New England had the widest negative disparities in the nation between our ranks in income and in charitable giving.  Originally we had no intention to publish it annually; that evolved because it had succeeded in generating broad and energetic discussion, and because we wanted to track any changes.   

As it grew in influence, and stimulated discussions around the issue of comparative generosity of states, we decided as a matter of policy to confine ourselves to the data alone, showing only self-evident arithmetical relationships between the numbers.  Presenting numbers in ascending or descending order does not interpret them; it simply makes them more interesting to people who want to interpret them—no one is driven to any particular conclusion by clarifying obvious arithmetical relationships between numbers.  Moreover, the numbers are of dollars, which are only one expression of philanthropic generosity on the part of donors.  So "confining ourselves to the data" means that we are talking only about numbers and dollars, which allude to, but are not the same as, generosity and donors.  Perhaps these distinctions are too subtle for most people, but they are real, they assist precision of thought, and therefore should be respected by scholars and professionals in the field. 

Our purpose is simply to stimulate discussions of charitable giving during the giving season, which before 1997 tended to be perfunctory and often negative—mainly warning donors to be careful.  We wanted to change the subject; these numbers did that.  We knew they would stimulate interpretations, and we thought it wiser to let others squabble over what the numbers mean than to entangle ourselves in necessarily speculative and inconclusive debates.  We were the first to state the clear limitations of the data, and in fact our first use of the phrase was clearly qualified and tentative: "a kind of crude but telling `Generosity Index'".   

The value of the GI has exceeded our intentions, pointing the way to new knowledge in a variety of applications.

 

—First, it showed that nationwide, charitable giving is not consistently related to income— the average state had a 20-point disparity (positive or negative, meaning a very broad range) between its income and its giving ranks.  This means we have no national culture of charitable giving—and come to think of it, why would we?  No national institutions exist to define or promote it.   

—Moreover, if giving is not determined by income, it must be culturally rather than economically conditioned.  This is enormously significant, because it means that giving can be improved through donor education—if it were based solely on income or economic factors, improving it would require improving income or those factors.  

— Then something intriguing happened: Governing magazine printed a map of the country with each state's rank on it, and color-coded for rank in groups of ten.  This was the first map of charitable giving nationwide, and it proved to be enormously interesting to many people.  It showed that contiguous states tend to have similar ranks on the GI, suggesting that the cultural influences on giving are significantly regional.  Subsequent maps were done by U.S. News and World Report, the Christian Science Monitor, and Barron's.  Generally these maps have made sense, coinciding with other knowledge we have about regional cultures, and have taught us new lessons:  

—They helped stimulate a correlation of giving with religious cultures, which helped to explain why Colorado and Wisconsin, both regional exceptions, were ranked lower (each had large Catholic populations, which characterize 7 of the bottom 8 states).  Independent corroboration of new knowledge tends to confirm its truth.   

—Finally, the GI has stimulated the promotion of philanthropy and charitable giving in a number of states, which is certainly a positive influence.