This is an article that appeared, in a couple of different forms, on previous versions of this blog. Since it critiques an opinion that’s become ‘common knowledge’ in some circles, I thought it would be good to repost it.
A few years ago, the image to the right started popping up on my Facebook feed. For those who can’t see images, here’s what it says: if churches paid taxes, it would generate enough revenue ($83.5 billion) to pay for the entire Supplemental Nutrition Assistance Program (SNAP) ($76 billion) and house everyone who is homeless.
The same concept later showed up in an article by Matthew Yglesias at Slate and, in turn, by Dylan Matthews at Wonkblog, who added material from an article by Ryan Cragun, Stephanie Yeager, and Desmond Vega titled “Research Report: How Secular Humanists (and Everyone Else) Subsidize Religion in the United States”. Both of these posts were picked up by Steve Benen at Maddowblog’s “This Week in God”.
Yglesias’s argument is fairly general, and there’s something to be said for at least a part of it. Yglesias points out that “discussing moral action is at the heart of many religious enterprises,” and that “much moral action plays itself out in the arena of politics.” It is somewhat perverse – in a soft sense – that a religious organization can advocate on behalf of the poor, but not on behalf of a partisan political party or candidate who also advocates on behalf of the poor. Likewise, it is somewhat perverse that a religious organization can organize against abortion, but not endorse political candidates who would work to end abortion legislatively.
The same issue applies to other nonprofit organizations as well, of course. Environmental nonprofits can advocate and organize to save our wetlands, but cannot endorse the candidates or parties who would actually protect those wetlands; housing nonprofits can work to increase affordable housing, but cannot endorse the candidates or parties who would create affordable housing trust funds; hunger nonprofits can operate food banks, but cannot endorse candidates or parties who would protect SNAP; and so on. If Yglesias’s argument applies to religious organizations, then it applies to all organizations that work on issues that have partisan implications. Yglesias’s argument isn’t so much against tax-exemption for religious organizations as it’s an argument against tax-exempt status more broadly.
What I want to focus on in this short series, though, is not the broader issue of tax exemption overall. Rather, I want to focus on the article by Cragun, Yeager, and Vega, as the estimates in this article seem to form the basis for assertions about how much additional revenue the government would have if religious organizations were taxed and all of the things that the government would do with it. As Matthews writes, “they’re not exactly disinterested parties; their research appeared in Free Inquiry, a publication of the Council for Secular Humanism.” However, I must disagree with Matthews’s assertion that “the data seems to check out.” The article by Cragun et al suffers from several major deficiencies:
I’ll take each of these on in greater detail below.
One final note before we dive into the article. Cragun et al refer repeatedly to the charitable status and subsidies that ‘religion’ enjoys or that ‘religions’ enjoy. ‘Religion’, of course, is an (often poorly defined) abstract concept, and abstract concepts – no matter how well defined – are neither given nonprofit status nor taxed. What Cragun et al are discussing in their article is not ‘religion’ or ‘religions’, but religious organizations. Moreover, they are not discussing all religious organizations – organizations that the authors consider both religious and predominantly charitable are not included – but a particular kind of religious organization: religious congregations. This is more or a pet peeve than anything else – the reification of religion often seems to lead to fuzzy thinking – but it is important to be clear about what we are discussing.
As I wrote in the introduction, one my pet peeves in the reification of religion. Imagining religion as something that exists separately from human beings or human institutions seems to lead almost inexorably to fuzzy thinking. Cragun et al are not discussing the taxation of religion or religions as abstract concepts. They are discussing the taxation of religious organizations. More specifically, they are discussing religious congregations, or those religious entities referred to by the Internal Revenue Service (IRS) as ‘churches’.
That being the case, two questions are reasonable. First, what is a church for the purposes of the tax exemptions that Cragun et al are discussing? Second, are the examples that Cragun et al use representative of those organizations?
The definition of ‘church’ used by the IRS is very broad and very fuzzy. The IRS uses a combination of attributes “generally attributed to churches” along with other facts and circumstances to determine whether any particular organization is a church. Moreover, the IRS does not evaluate “the content of whatever doctrine a particular organization claims is religious,” meaning that there is not and cannot be any legal requirement for a church to believe in deities, miracles, or anything else supernatural.((Internal Revenue Service, Publication 1828: Tax Guide for Churches and Religious Organizations, 27)) In fact, there are secular organizations organized as churches that qualify for tax exemption under that rubric. The important point here is that what a church is under IRS guidelines does not necessarily correspond to what a church is in popular usage.
When it comes to examples, Cragun et al use a very narrow set of traditionally religious – and entirely Christian – organizations.((Cragun et al refer to several specific religious groups and congregations which come from these traditions: Baptists, The Church of Jesus Christ of Latter-Day Saints (Mormons), the United Methodist Church, Episcopalians, Pentecostals, and Presbyterians. For several of these traditions, the specific kinds of churches are left undescribed. While there is room for debate over how to describe Mormonism, there is no doubt that it fits better with Christianity than with any other broader tradition.)) While it makes obvious sense to focus on Christianity as it is the most popular single religious tradition in the United States it is important to note that the authors’ treatment of ‘religion’ does not reflect the incredible diversity found in this country. Some of this diversity is reflected in structural differences between churches that would have effects on taxes were they imposed: does a tradition make use of parsonages or not? do its ‘churches’ generally own buildings, rent space, or meet in houses? if its ‘churches’ do have buildings, what kind of buildings are they? and so on.
All of this may seem unnecessary, but it is important to be clear about what Cragun et al are actually discussing in this article. While they claim to be writing about how ‘religion’ is subsidized in the United States, they are in fact looking only at a subset of religious organizations that are designated as ‘churches’ under federal tax law. And when it comes to examples, they are working with an even smaller subset: ‘churches’ within the Christian tradition. They are not discussing religious organizations – including hospitals, homeless shelters, soup kitchens, etc. – that are not churches or integrated auxiliaries.
Cragun et al spend the first major section of their article mounting a preemptive defense against what they see as a potential criticism of their work:
By suggesting that these groups should pay taxes, we are likely to be criticized by those who think that religions are largely charitable institutions engaged in beneficial service or charitable work and should therefore be exempt from taxes.((Cragen et al, 39))
As we will see, being charitable is only one of many possible reasons that an organization might be tax exempt; there are many organizations – both religious congregations and others – that are tax exempt for reasons other than their being charitable. However, because Cragun et al effectively treat being charitable as the only criterion for tax exemption, it’s worth considering the understandings of charity that they offer. Cragun et al provide three descriptions of charity.
First, Cragun et al describe charity as involving “‘serving people’s physical needs’ (feeding and clothing the poor, building schools, and the like).” Occasionally, this slips into the more specific “physical needs of the poor.”((Cragun et al, 39-40))
This is as good a description as any and certainly reflects the popular concept of charity. It does not, however, reflect the diverse array of organizations that are tax exempt. Even the parenthetical provided by Cragun et al reveals the separation of this understanding of charity from the broader legal framework of tax exempt organizations: the acts of feeding and clothing the poor certainly serve people’s physical needs, there can be reasonable debate over whether building a school does so; such debates would also exist around symphonies, theaters, animal shelters, and many other organizations. This description may be adequate for considering the relatively narrow field of charitable organizations, but it is inadequate when considering all of the kinds of organizations that are tax exempt.
Second, Cragun et al describe charity as involving the “addressing people’s ‘spiritual concerns.’”((Cragun et al, 39-40)) While Cragun et al seem to accept the idea that addressing people’s spiritual concerns might be considered charitable, they reject the idea that religious congregations address those concerns charitably for an interesting reason: pastors and other religious functionaries are paid to address those concerns: “It is no more ‘charity’ than a college professor teaching a class or a social worker helping a family is charity.”((Cragun et al, 40))
This description of charity confuses individuals with organizations. It is absurd to suppose, as Cragun et al appear to do, that the charitable status of an organization is dependent upon whether the people working with it are volunteers or employees. I suspect that Cragun et al are aware of this absurdity. It is inconceivable, for example, that Cragun et al would consider the American Red Cross – an organization that they use as an example of an efficient charitable organization((Cragun et al, 40)) – to not be a charitable organization because it has paid employees, some of whom are remunerated quite handsomely.((See the American Red Cross’s 2012 Form 990 for a list of the highest compensated employees.)) Many other charitable organizations also make use of paid employees. Whether an individual is acting charitably or not when working with an organization – whether, that is, she is acting as a volunteer or an employee – has no relevance when it comes to the charitable status of the organization.
Third, Cragun et al describe charity – or, at least, a criterion “of what it means to be a charitable organization for tax purposes” – this way: charitable organizations must provides goods or services the discontinuation of which would “result in a legal requirement for public funds to continue the function.”((Cragun et al, 40))
This criterion is absurd on its face. While it is true that one generally accepted purpose of charity is to lessen the burden on the government,((Exempt Purposes – Internal Revenue Code 501(c)(3) )) there is no legal requirement that public funds be used, for example, to feed people whose local food pantry has closed. Indeed, many food pantries operate precisely to fill the gap left by the Supplemental Nutrition Assistance Program. Similarly, there is no legal requirement that the government provide housing for those whose housing needs are not met by homeless shelters. The same is true for the vast array of other kinds of charitable organizations: while they may reduce burdens on the government, they are not a replacement for government services and the government is under no obligation to fill their role should they cease to exist. The idea that such a legal requirement is part of what it means to be a charitable organization for tax purposes – at least in the sense that Cragun et al seem to be using the phrase – is simply false.
The descriptions provided by Cragun et al do not account for the wide variety of nonprofit tax-exempt organizations in the United States. For example, they don’t account for educational institutions, arts organizations, animal shelters, or scientific research organizations. I think all of us – Cragun et al included – would say that these kinds of organizations are valuable even if they don’t serve people’s physical needs and even if we may disagree with some of the work that they do. Fortunately, the legal framework under which these and other tax exempt organizations operate is not that of the relatively narrow ‘charity’, but of the broader ‘nonprofit’.
The most well-known nonprofit designation – and the designation under which churches and other religious congregations fall – are 501(c)(3) organizations, so-called because their tax exempt status is governed by 26 US Code § 501 subsection (c) paragraph (3). This paragraph lists a set of kinds of organizations that are tax exempt:
Corporations, and any community chest, fund, or foundation, organized and operated exclusively for religious, charitable, scientific, testing for public safety, literary, or educational purposes, or to foster national or international amateur sports competition (but only if no part of its activities involve the provision of athletic facilities or equipment), or for the prevention of cruelty to children or animals, no part of the net earnings of which inures to the benefit of any private shareholder or individual, no substantial part of the activities of which is carrying on propaganda, or otherwise attempting, to influence legislation (except as otherwise provided in subsection (h)), and which does not participate in, or intervene in (including the publishing or distributing of statements), any political campaign on behalf of (or in opposition to) any candidate for public office.
As you can see, charitable organizations are included in this paragraph. But so are a host of other kinds of organizations that may or may not qualify as ‘charitable’ in the public imagination, including religious organizations. Churches, in other words, are not tax exempt because they are charitable – though they may be charitable – but because they are religious.((The question of whether tax exemption for organizations on account of their status as religious organizations is constitutional is taken up by the Supreme Court in Walz v. Tax Commission of City of New York, 397 US 664 (1970). The court found that such exemption did not violate the first amendment’s prohibition against the establishment of the religion for two reasons. First, the law – an exemption from property taxes in this case – had not “singled out one particular church or religious group or even churches as such; rather, it has granted exemption to all houses of religious worship within a broad class of property owned by nonprofit, quasi-public corporations which include hospitals, libraries, playgrounds, scientific, professional, historical, and patriotic groups.” Second, the taxation of religious congregations would result in greater entanglement between the state and religious congregations than not taxing them: “Elimination of exemption would tend to expand the involvement of government by giving rise to tax valuation of church property, tax liens, tax foreclosures, and the direct confrontations and conflicts that follow in the train of those legal processes.” The point here is not only that tax exemptions for religious congregations – on account of their status as religious organizations – has been found constitutional, but that balancing the demands of the two religion clauses of the first amendment are often rather complicated.)) The larger framework created by this paragraph is vital to a number of different kinds of nonprofit organizations.
Moreover, the nonprofit framework is even larger: many other kinds of organizations are governed by other paragraphs of 26 US Code § 501(c). For example, Cragun et al suggest that one compromise between the charitable activities of religious congregations and their other activities would be to allow the charitable activities to be tax exempt while treating the congregations themselves as “civic leagues or sports clubs or any other volunteer organization that exists for entertainment or the benefit of its members,” which they claim are not tax exempt.((Cragun et al, 44)) Depending on how such congregations are reorganized, however, they might qualify for tax exemption under § 501(c)(4) (civic leagues or organizations not organized for profit but operated exclusively for the promotion of social welfare), § 501(c)(7) (clubs organized for pleasure, recreation, and other nonprofitable purposes, substantially all of the activities of which are for such purposes and no part of the net earnings of which inures to the benefit of any private shareholder), or even other parts of § 501(c)(3) (what is a Bible study group other than a literary society, after all?). Creative religious congregations could reorganize themselves as a variety of different organizations each covered under a different portion of § 501(c).
This is all, of course, rather preliminary. The meat of the article is found in the estimates that Cragun et al make regarding the subsidies granted to religious congregations through tax exemption. However, the points I’ve provided here are valuable for two reasons.
First, Cragun et al overstate the range of their estimates. They are not estimating what subsidies ‘religion’ receives through tax exemptions, but only what subsidies religious congregations receive. In their proposed ‘compromise’, they narrow the field even more as they suggest separating the charitable activities of those congregations from other activities. Since many religious people and organizations engage in charitable – and other tax exempt – activities for religious reasons, this parsing out of ‘the religious’ and ‘the charitable’ is gravely problematic. Is a food pantry run by a religious congregations an act of charity or an act of religion? Is the addressing of spiritual concerns – by a volunteer if that would make Cragun et al feel better – a charitable act or a religious one? Parsing out such distinctions is not only a difficult task, but asking the government to perform this analysis would necessarily create far more entanglement between church and state than anyone should be comfortable with.
Second, the way that Cragun et al make use of ‘charity’ is overly narrow and the application of such use would have ramifications far beyond religious congregations. If the only tax exempt organizations were those (1) that address people’s physical needs, (2) that are staffed entirely by volunteers, and (3) the discontinuation of which would “result in a legal requirement for public funds to continue the function,” then most nonprofit organizations would lose their tax exempt status. Organizations losing their tax exempt status would include the American Red Cross (which Cragun et al use as an example of an efficient charitable organizations) and the University of Tampa (at which Cragun teaches and at which his coauthors were students when the article was written). Such loss for those two organizations might be a small matter, however, as the list of organizations would also include almost every other nonprofit. The authors’ view of what constitutes ‘charity’ and what should allow for tax exemption is simply not tenable.
In the end, the Cragun et al simply appear ignorant about the broader nonprofit framework in which religious congregations operate. Were they attempting only to provide estimates for the value of the subsidies that religious congregations receive, this would be a trifling detail. Since they also spend about a third of article implying that religious congregations should not be tax exempt, however, and since they make factual claims about tax exempt organizations that are dubious at best, their ignorance is a matter of some importance and reflects poorly on their ‘research’.
Before I turn to the estimates that Cragun et al make, I want to address a couple of paragraphs from earlier in the article. In these paragraphs, Cragun et al compare the charitable giving – defined here by Cragun et al as work that “addresses the physical needs of the poor” – of a few different organizations: the Church of Jesus Christ of Latter-Day Saints (LDS), the United Methodist Church (UMC), Walmart, and the American Red Cross. The figures that they offer are: $1 billion in charitable aid from the LDS over twenty-three years or 0.7% of annual revenue per year, $62 million from the UMC in 2010 or 29% of revenues, $1.75 billion from Walmart ‘each year’, and 92.1% of revenues from the American Red Cross.((Cragun et al, 40)) The figures provided in these two paragraphs are problematic for several reasons and reveal some severe deficits in the authors’ reasoning.
First, these are incongruous figures. We are presented with an absolute amount and a percentage of revenue for the LDS over twenty-three years, an absolute figure and a percentage for the UMC in a single year, an absolute figure for Walmart in what is presented as a ‘normal’ year, and a percentage of revenue for the American Red Cross in an unspecified year.These figures are presented as though they are comparable, but the variation in them means that they are not. If we want to compare the amounts given to charitable causes by different organizations, then we want to see either absolute numbers or percentages of revenue – preferably percentages of revenue – for the organizations over similar periods of time.
The reason we would prefer to see percentages is because of the different sizes of organizations. This is especially true of the organizations that are at play here. To see this, let’s take a single year’s numbers:
Two notes must be made immediately here. First, Walmart’s figure for charitable giving in 2010 is significantly lower than the figure given by Cragun et al for their charitable giving ‘each year’. Second, the figure given for the American Red Cross is for program expenses, which some readers will note is not the same as money spent “directly addressing the physical needs of those it intends to help.”((Cragun et al, 40)) Both of these issues will be discussed in more detail below.
For now, notice the significant differences in scale. Walmart’s net sales are almost 700 times the total revenues of the LDS and almost 2000 times the total revenues of the UMC. And yet in absolute terms their giving is a mere fourteen times larger than that of the LDS and slightly more than ten times that of the UMC. More tellingly, in terms of percentage of net sales or total revenue, Walmart’s giving is less than a quarter of that of the LDS and about one twentieth that of the UMC. The point here is not to shame Walmart, but to show that comparing the giving of two organizations whose total revenues are separated by several orders of magnitude in absolute dollars may give a misleading impression. This is all the more true when one of those organizations is one of the largest and wealthiest multinational corporations in the history of the world.
Second, the statement that Walmart “gives about $1.75 billion in food aid to charities each year”((Cragun et al, 40)) is simply false. As we saw above, in 2010, Walmart’s charitable giving was considerably less. I was unable to locate a reference for the figure given by Cragun et al (their citation leaves something to be desired), but the link that Cragun et al provide indicates that in 2013 Walmart gave $1.3 billion in cash and in-kind donations globally. Their 2013 Annual Report states that “Walmart and the Walmart Foundation’s charitable contributions surpassed $1 billion in cash and in-kind donations,”((Walmart, 2013 Annual Report, 3)) indicating that this is likely the first year that Walmart exceeded $1 billion in giving. Walmart’s 2012 Global Responsibility Report indicates that in that year they gave slightly less than $1 billion.((Walmart, 2012 Global Responsibility Report, 14)) While it is certainly possible that Walmart has given $1.75 billion in food aid at some point, the statement that they give this amount ‘each year’ strikes me as misleading.
Third, the statement that the American Red Cross spends 92.1% of its revenue “directly addressing the physical needs of those it intends to help”((Cragun et al, 40)) is also false. Cragun et al got their figure from the American Red Cross’s summary on Charity Navigator,((This summary is regularly updated, so the current figure listed there may not match the figure that Cragun et al use.)) which provides the percentage that the organization spends on ‘program expenses’. As some readers will recognize, ‘program expenses’ is a category of expenses from the Form 990 and contains all money spent on programs, not only money spent directly addressing physical needs. For example, the American Red Cross’s 2011 Form 990 – which details their finances from 2010 – includes more than $1.5 billion spent on salaries, pensions, other benefits, and payroll taxes;((Remember that Cragun et al earlier argued that if people are paid to provide services, those services are not charitable (Cragun et al, 40). )) as well as a variety of other expenses that may or may not qualify as direct physical assistance.((American Red Cross 2011 Form 990, 11))
As nonprofit professionals know, measuring the direct impact of an organization is exceedingly difficult and ‘program expenses’ can be a useful proxy. But program expenses are decidedly not the same as money spent ‘directly addressing the physical needs of those an organization intends to help’.
In these two paragraphs, Cragun et al present the inaccurate results of incomplete research on Walmart and the American Red Cross. Moreover, they present those results in a misleading way. As we will see, this is characteristic of how the authors make their estimates.
When Cragun et al turn to estimating the subsidy that religious congregations receive through being exempt from federal income taxes, they use a simple formula. They begin with the total amount given to religious congregations in 2009, according to Giving USA: $100.95 billion.((Cragun et al, 41)) They then assume that all religious congregations would fall into the highest federal corporate income tax bracket of 35%.((Cragun et al, 42)) From here, the math is easy: 35% of $100.95 billion is about $35.3 billion, which is the estimate that Cragun et al arrive at.((Cragun et al, 42))
There are multiple problems with this approach.
First, the assumption that all religious congregations would fall into the highest federal corporate income tax bracket cannot be accurate. The highest tax bracket – with a rate of 35% – requires that a corporation have a taxable income of at least roughly $18.3 million. The idea that most religious congregations would have taxable incomes of that amount is absurd on its face. Moreover, it is incompatible with other assumptions that Cragun et al make in the article. For example, when estimating the total property value of property owned by religious congregations, they assume that there are 335,000 congregations in the United States.((Cragun et al, 42)) If just half of those congregations had taxable incomes of $18.3 million – with the others having taxable incomes of just $1 each – then the total amount given to religious congregations could not be $100.95 billion, it would have to be at least $3 trillion, or almost thirty times the amount that Cragun et al use in their estimates.
Using the initial estimate for total giving to religious congregations – $100.95 billion – and the estimated number of congregations that Cragun et al use, we would find that the average revenue of a congregation cannot be more than $302,000 (about 1.7% of what Cragun et al estimate). This would put congregations into a significantly lower tax bracket applying to corporations with taxable incomes from $100,000 to $335,000 and paying $22,250 plus 39% of the amount over $100,000. This would mean an average tax of just over $101,000 or about 33.5% of taxable income. This would result in a total federal income tax bill of $33.8 billion.
That difference is trivial perhaps. My estimate is about 96% of the estimate provided by Cragun et al. The second problem is much larger: corporations do not pay taxes on their total revenue, but on their taxable income. Just like individuals can take deductions from their gross income to calculate their lower taxable income, corporations can take deductions from their total income to do the same: salaries and wages, pension plans, employee benefit programs, maintenance, bad debts, rents, licenses, depreciation, and advertising are all tax deductible within certain limits. This means that the taxable income of a religious congregation would likely be substantially lower than its total revenue.
Again, we can make some rough estimates using information in the article itself. Earlier in the article, Cragun et al point out that “on average, ‘operating expenses’ totaled 71 percent of all the expenditures of religions, much of that going to pay ministers’ salaries.”((Cragun et al, 40)) If we make a very charitable assumption in favor of Cragun et al that a mere 35% of ‘all the expenditures of religions’ go to pay ministerial salaries and that this is the only deduction that a congregation might take, that would reduce the taxable income the average congregation from $302,000 to about $196,000 and thus their tax burden from $101,000 to about $60,000 or about 20% of total revenue. This would result in a total federal income tax bill of about $20.2 billion or about 57% of what Cragun et al estimate. Of course, religious congregations would be able to take many other deductions, lowering their taxable incomes even more.((In principle, of course, these deductions would still be subsidies to religious congregations, but only in the sense that they are subsidies provided to every corporation.))
These deductions – along with some other tax tricks – are the reason that economists usually don’t use the statutory corporate tax rate. Instead, they use the effective corporate tax rate, which in 2010 was 12.6%. Using the total amount donated to religious organizations, this yields a total of about $12.7 billion, or a mere 36% of the estimate made by Cragun et al.
We can see here the issues raised in the previous section. The estimates provided by Cragun et al are presented as accurate while based on a woefully incomplete understanding of how corporate taxes actually work. This is, first of all, misleading. More problematically, however, when we look at the methods that Cragun et al use to make their estimates we see no reason to suppose that those estimates would be accurate.
When it comes to state income taxes, Cragun et al relegate their methods to a footnote: they break down the total amount donated to religious organizations into the amount donated per person, use state populations to estimate the amount donated to religious organizations per state, and then use state-level corporate income tax rates to calculate the amount that each state would receive in taxes if religious congregations were taxed.((Cragun et al, 45)) The figure that they arrive at is $6.18 billion.
Unfortunately, that isn’t enough information to reconstruct their estimate. It sounds like they took the total amount donated to religious organizations ($100.95 billion) and divided that by the population of the United States (about 309,000,000 in 2010) to arrive at the amount donated per person (about $32.66), used state populations to estimate an amount donated per state, and then multiplied that amount by the appropriate corporate tax rate in each state. However, this yields a total tax subsidy of only about $624 million. A far better method would be to take the number of religious congregations in each state (provided by the Association of Religion Data Archives), multiply that by the average revenue of each congregation to arrive at a total congregational revenue per state, and then multiply that by the appropriate tax rate for each state. Obviously, we can’t use the average congregational revenue provided by Cragun et al ($18.3 million) as that would give far too high an estimate. If we use the more realistic figure of $302,000, however, we end up with an estimate of about $6.3 billion, slightly higher than the estimate given by Cragun et al.
As we can see, there is no reason to expect the method – as Cragun et al describe it – to provide an accurate estimate. Besides the continuing problem of tax deductions, there are two substantial issues.
First, corporate taxes aren’t based on the spending habits of individuals, but the earnings of the corporations being taxed.
Second, the way that Cragun et al claim to break down giving and use state population presumes an equal rate of religiosity in each state. But this isn’t true. Vermont, for example, is the least religious state, with only about 44% of the population being moderately or very religious; Mississippi is the most religious, with about 91% of the population being moderately or very religious. We can safely assume that giving to religious congregations is correlated with religiosity, not simply with population. The estimates made by Cragun et al are therefore likely to be wrong even at the state level.
That leaves us, as above, in the position of not having any reason to believe that the methods employed by Cragun et al would provide accurate estimates.
I want to briefly highlight one additional area that Cragun et al address: property taxes. In a single paragraph, Cragun et al make a variety of errors.
First, they bring in their estimate of the number of religious congregations that caused such problems for their estimate of the revenue of “most local congregations.” ((Cragun et al, 42)) This is fair enough, but the Hartford Institute for Religious Research, from which they get their estimate, refers to the U.S. Religious Census, which in turn refers to the Association for Religion Data Archives (ARDA). I already used some data from ARDA above, and it will be important again in a moment.
Second, Cragun et al estimate the average value of congregational property by looking at a sample of 47 properties that (1) all belong to congregations in the broader Christian tradition and (2) are all in the Tampa, Florida, area. While it may be true, as Cragun et al claim, that “property values in Tampa are close to the national average,”((Cragun et al, 45)) this is a woefully small and homogenous sample.
Third, Cragun et al average “the total number of churches across all fifty states.”((Cragun et al, 42)) This is where the ARDA data would prove useful as, of course, there are not an equal number of congregations in each state. While Cragun et al appear to assume that there are 6,700 congregations in each state, the truth is that the number of congregations varies considerably. Texas, for example, has more than 27,800 congregations, while Rhode Island has fewer than 700. Taking the most and least religious states I referred to earlier as examples, Vermont has only about 850 congregations and Mississippi has about 6,700. Given that “property taxes are paid at the state level”((Cragun et al, 42)) – which isn’t actually true, but state medians can serve as a proxy for the complex jurisdictions in which property taxes are paid – having the appropriate number of congregations in each state might be important.
This method also, of course, raises a further question. If in the case of property taxes Cragun et al could use an estimate of the number of congregations in each state and the value of their property and the state property tax, why did they not use a similar method when calculating state income taxes?
Once again, regardless of the amount actually estimated, we are left with no reason to suppose that this method – which assumes an equal number of congregations per state and estimates the average value of congregational property using such a small sample – would produce an accurate estimate.
As Cragun et al sum up their estimates, they reach the conclusion that subsidies to religious organizations total to $71 billion plus the totals of estimates that they were unable to make. If nothing else in the article raises eyebrows, this should. Cragun et al estimate that religious organizations receive a total of $100.95 billion each year and, if taxed, would pay $71 billion; that’s a tax rate of 70%, which is absurd on its face.((This actually reveals another methodological problem: taxes don’t stack like this. Money paid in certain taxes is deductible from other taxes, so simply summing up all of the estimates will result in a bloated figure.))
This leaves us with a couple of options. First, we can accept – as I’ve proposed here – that the methods that Cragun et al used to arrive at their estimates are faulty and their estimates are likely wrong. Second, we could defend Cragun et al by using a broad understanding of ‘subsidies’. Tax deductions, after all, can be understood as subsidies to those entities that can take them. This defense, however, would make the entire article misleading. Cragun et al would not be estimating the amount that religious congregations would pay in taxes “if they were required to pay taxes the same as for-profit corporations do,”((Cragun et al, 39)) but the amount that religious congregations would pay in taxes if they were liable for the same taxes as for-profit corporations and did not receive the same subsidies as those corporations.
And so we are left in the position we have been in: we have no reason to suppose that the methods used by Cragun et al would result in accurate estimates and we have many reasons to suppose that they would result in inaccurate estimates.
As we’ve seen above, “Research Report: How Secular Humanists (and Everyone Else) Subsidize Religion in the United States” by Ryan Cragun, Stephanie Yeager, and Desmond Vega has multiple issues:
This is not, of course, a complete list of issues with this article. Many more posts could be written on the inaccuracies contained within it. But these items should be enough to call into question the conclusions reached by Cragun et al.
I won’t attempt to psychologize these issues, but I will note – along with Dylan Matthews, who used the article in a post on Wonkblog – that Cragun et al are “not exactly disinterested parties; their research appeared in Free Inquiry, a publication of the Council for Secular Humanism.” That authors with ideological investments sometimes make poor arguments in favor of those investments is not news. We all do it now and again.
I am far less concerned with the fact that an poorly argued article appeared in Free Inquiry than I am with the fact that the article was used as the basis for posts put forth in a variety of other settings. That the same Dylan Matthews who knew that recognized that this was not disinterested research was also able to write that “the data seems to check out” only indicates that he couldn’t have actually checked out the data. That the claims made in the article were credulously picked up by otherwise reasonably skeptical media only goes to show that any media outlet can be subject to confirmation bias.