The Contradiction of Compassion: Private Charity and the "Solution" to Dependence

Stephen T. Ziliak, University of Iowa


Following a sharp reduction in the amount of public outdoor relief going to the poor in the late 19th-century, the private charities initiating the reduction became well-financed and generous to the "worthy" poor. They approximated the public expenditures per relieved household that had obtained during the spree which motivated the movement. And yet to the extent that labor supplies were elastic, the induced generosity thwarted the reformers' objective of increasing self-reliance among the "worthy."

As a strategy for encouraging self-reliance, in other words, the privatization of poor relief did little better in the United States than public outdoor relief, before and after the emergence of the welfare state. Exits from the rolls by way of higher earnings occurred at the same rate as those from the Aid to Families with Dependent Children program. In fact, the average duration of a spell on poor relief has changed little between the 1820s and the 1990s. The privatization of poor relief in the late 19th-century was not like Beecham's pills, "working like magic" to uplift weak character and human capital. The findings may surprise observers during the nth revival of a dependency problem. But the resolution of an old contradiction in price theory predicts the futility of privatization as a means of increasing self-reliance.

The Abolition of "Welfare" in Historical Perspective

The most ambitious attempt to privatize poor relief in America happened between the early 1870s and the 1900s, led by the Charity Organization Society. Ten of the largest cities abolished public outdoor relief - the relief one consumes outside an asylum - and many others, such as Indianapolis, sharply reduced theirs (Almy 1899, p. 31). In Brooklyn, for example, the annual cost of public outdoor relief in 1877 was 10 or 12 dollars per household. The following year the payments were cut by over half, and were then abolished in 1879, directly affecting more than 46,000 people (Devine 1899, p. 188).

Josephine Shaw Lowell (1884) and Edward T. Devine (1899), both leading charity organizers, examined in Brooklyn ten years of aggregate data on indoor relief and on private outdoor relief supplied by organized charity. Finding no appreciable increase in expenditures or in caseloads after abolition, they concluded, 15 years apart, that "the abolition of the outdoor relief in Brooklyn resulted in a real improvement in the status of the very poor" (Devine, p. 188). "There could scarcely be a stronger proof that the stopping of out-door relief does not cause the suffering that is anticipated . . . the need supplied by public out-door relief is in fact created by it" (Lowell, p. 62). Apparent success in Brooklyn and then in Philadelphia inspired the Indianapolis Charity Organization Society to lobby for abolition (Associated Charities 1886, pp. 46-47).

Yet the "real improvement" was not properly demonstrated, despite Lowell's declaration of "proof." Finding little increase in indoor relief or in one or two organized charities after abolition does not demonstrate that the wages and employment of would-be recipients of public outdoor relief had increased. The "improvement" hypothesis has low power against important alternatives, such as death by starvation. After all, Sir Charles Trevelyan, a leading organizer in England, regarded "deaths by starvation as a discipline, a painful one admittedly, but nevertheless a discipline" (Hart 1960, p. 99). The American and English realities were not in fact so horrific. But the data examined by Lowell and Devine suggest little about the "creation" of "need," and suggest nothing about material well-being and the allocation of time after abolition. Yet evidence from Indianapolis and from metropolitan England (Humphreys 1993, p. 16) suggests that 60-70% of the 46,000 recipients of public outdoor relief in Brooklyn would have been "worthy" of private outdoor relief, and would have accepted it. Thousands have been mislaid. Still, in The Emergence of Modern America Allan Nevins all but claimed as his own the written word of reformers. "Official provision for the poor, . . ., was rude, unsystematic and wretchedly administered. One great city after another furnished outdoor relief on the most lavish and careless scale," wrote Nevins, "yet to the surprise of observers no increased demand fell upon private or public relief agencies [following abolition in Brooklyn], and no unusual suffering appeared" (1927 [1969], p. 327).

The privatization movement has in recent historiography both foes and friends. But here the history of ideas looms largely instead of evidence on the history of economic behavior; in some cases, appeals to the history of ideas have been posed as substitutes for empirical investigation. Alice Kessler-Harris (1981) wrote in an early book on women's labor history that "[a] traditional charitable approach demanded care for the poor - clearly part of women's work. Yet public policy in the post-Civil War years rejected the notion of charity" (p. 112). Kessler-Harris claimed that "social Darwinism" (Hofstadter 1944 [1955], pp. 3-12) was the guiding philosophy. "The fit must refrain from overt help to the poor lest they disrupt society's natural laws . . . Money and food were not to be given out as they threatened to destroy character" (Kessler-Harris, p. 112).

The social histories by Michael Katz (1986, Chp. 3) and by David Montgomery (1993, Chp. 2) found little to recommend, and on similar grounds. Katz, for instance, argued that the "method [of being helpful] taught dependence, because only an outward show of deference merited relief . . . In fact, increased dependence became the price of continued support. In the end, the Charity Organization Society as much as the villains of their drama, purveyors of indiscriminate charity, taught the poor to be paupers" (Katz, pp. 67-68). "Charity organization failed on every major dimension" (p. 80).

Gertrude Himmelfarb, by contrast, has essentially called for a revival of the movement. Himmelfarb would like to follow closely the letter of the New Poor Law (Himmelfarb 1991, pp. 189-190; Himmelfarb 1995b; Webb and Webb 1909; McCloskey 1973; Boyer 1990). Himmelfarb speculates that a "solution to the problems that beset us" is a Victorian attack on "moral pathologies" like "welfare dependence," promising a "fighting chance" of undoing the "culture of dependency" created by the welfare state (Himmelfarb, Wall Street Journal, Feb. 7, 1995a). "If we seriously want to discourage dependency, criminality, . . ., and the like," says Himmelfarb, "we have to do more than create a more rational system of economic and legal incentives . . . We have, in short, to relegitimize morality as the basis of social policy and restore the language of virtue and vice."

While Mary MacKinnon (1987) and Robert Humphreys (1993) have studied quantitative aspects of the crusade in England, economists have not reconsidered the reforms in the United States, let alone at the micro-level. But analyzing micro-data here and there will not permit a new rank-ordering of the various scientific achievements. Although it has enjoyed continuous usage since the 18th century, the economic theory of poor relief is in grave condition.

The Contradiction of Compassion: A Model of Futility

The problem with the economics of abolition is that it ignores the long-run effects on the labor market produced by the "crowding-in" of private charity. Private charity has been tacitly or expressly described by observers on the left and on the right as ephemeral, an immaterial addition to the budget of a poor person. Liberal and socialist scholarship since Dickens has figured the middling classes a crowd of Scrooges before Marley's Ghost. Tiny Tim and his able-bodied family suffer from the stinginess, which is taken to be a prima facie case for governmental intervention. On the other hand, conservative scholarship has elevated the Ladies Bountiful and the Andrew Carnegies, who, when left alone, share generously in their bounty, but whose beneficence does not produce negative effects on labor supply.

And yet private charity cannot be in one time and place both small and harmful and large and unharmful (see McCloskey 1973, pp. 419-427, for a similar point). Modern economists (and humanists) are telling both stories. But each story harbors a fundamental contradiction. The problem is that theories of work disincentives and of the finance of poor relief are divorced. And yet the contradiction can be observed in a familiar enthymeme:

"Public outdoor relief reduces labor supply and encourages dependence;"

"Public outdoor relief largely crowds-out private charity;"

"Therefore, abolishing public outdoor relief increases labor supply

and encourages self-reliance."

The enthymeme is a reasonable description of beliefs held by eminent economists - among them, nowadays, Milton Friedman - and by scholars such as Charles Murray and Gertrude Himmelfarb. Charity organizers crafted the enthymeme into a tool for policy reform (ICOS 1886, pp. 46-47; Almy 1899; Devine 1899). To be sure, the explicit premises are in isolation sound pieces of price theory. But price theory suggests that all income subsidies, whether they are financed publicly or privately, reduce labor supply and encourage dependence. Despite a long history of practice, the two premises cannot be properly viewed in long-run analysis in isolation. The usual analysis does not produce an equilibrium. The usual analysis implies that when public relief is abolished, private charity is crowded-in largely - in Roberts' model, dollar-for-dollar - and yet the private transfers are said to be immaterial to the budgets of the recipients. This is the contradiction of compassion.

But of course when private charity is crowded-in, the monies must find their way into the hands of the poor, whose unemployment and destitution are seen by others to be worthy of help, but whose industriousness and well-being are put into question. The induced generosity may line the pockets of a preacher and a reformer instead of the pockets of a ward boss, but there are no theoretical reasons for thinking that the material standards of relieving the "worthy" change substantively in the context of institutional replacement. Donor demand for the public good may during replacement shift downward, adjusting donation behavior to reflect the change in attitude about "worthiness," leaving expenditures per relieved household unchanged. The resolution of the contradiction is:

"Therefore, abolishing public outdoor relief crowds-in private charity,

which hardly increases labor supply or promotes self-reliance,"

so long as leisure is a normal good among the "worthy," among those who receive help. As a first approximation, then, an economics without personal relations, when public outdoor relief is sharply reduced, private charity is crowded-in, and the average length of a spell on poor relief is not much diminished. The simple inversion of Roberts' model into a model of crowding-in, when joined with the always obvious implication that subsidies of time away from wage labor increase its demand, takes one a long way toward explaining data on the first privatization movement.

The Indianapolis Experiment

There is a confusion in the literature concerning the measurement of crowding-in and crowding-out, a confusion which continues to exert a large effect on beliefs about the history of charity. Richard Steinberg surveyed the econometric literature on crowding-out, finding regression coefficients of size -$0.05 to -$0.30 in 9 cross-sections and time-series since the Second World War (Steinberg 1989, p. 150). Steinberg concluded from the survey "that when a local government reduces its provision of social services, it cannot count on the nonprofit sector to replace a very large part of these services" (Steinberg 1989, p. 155). In an otherwise fine paper, Steinberg's conclusion, shared by many economists, is too cavalier; it is at least inaccurate as a point of historical fact. After all, the studies cited did not estimate crowding-in effects. That is, the samples were not drawn to measure the response of private charity after a sharp reduction in expenditures on public relief. The symmetry of crowding-in and crowding-out found in the models has been assumed to be a characteristic of actual economies, but the assumption has not been investigated.

The studies did not estimate the size of crowding-out following a major institutional change, and yet the theories of crowding, including Steinberg's (1987, 1989), are relevant only in a context of major institutional change. Roberts is explicit on the point: "the model. . . predicts that private charity first became negligible when government first intervened in a significant way in the charity market" (Roberts, p. 142). While Roberts is mistaken in suggesting that the 1930s marked the first significant intervention - after all, observers in the 1870s, economists among them, thought the interventions warranted a privatization movement - the implication for empirical investigations is clear. And yet the widely cited coefficient of -$0.30 reported by Abrams and Schmitz (1984, p. 566) was found in a nation-wide cross-section of private donations and state and local welfare payments in 1979, a year of ordinary activity.

Indianapolis was chosen as the place of study because it likely provides an upper bound on the labor market success of the privatization movement, an upper bound on the speed and scale of the transitions of the "worthy" poor from dependence to self-reliance following a major institutional change. Public outdoor relief was not abolished in Indianapolis but it was reduced to a low level, falling in real terms in 4 years from $50,000 to less than $7,000. Public outdoor relief per relieved household was in 1877 equal to about 3 weeks the wages of an able-bodied common laborer. Expenditures per relieved household were by 1885 16% the earlier level.

Charity was not dead as a door-nail, to begin with. In 1876 Indianapolis supported twelve private charities. But a dozen more charities were established between 1877 and 1885, and between 1885 and 1895 sprang 20 more. A sample of the names suggest the diversity of services offered and the populations served: the German Ladies Benevolent Society, the Ladies' Hebrew Benevolent Society, the St. Boniface Benevolent Society, the Socialistic Sick Benefit Society, the Alpha Home for Aged Colored Women, the Woman's Relief Corp, and the German Lutheran Orphan's Home.

The middle class and upper class donations to the "scientific" charities did not alone replace the 1876 public expenditures, but clearly poor relief was in the early 1880s largely financed and administered by voluntary efforts. Accounting for church-based charity and other charity to the very poor may reveal approximate replacement. [Full replacement in total expenditure would be an odd event, however, given the middling enthusiasm for desert as an ethical category, even, if only weakly preferred, among Catholics (Katz, Chp. 3). It was in fact the conflation of need and desert as ethical categories which justified the policy of helping only those who after thorough investigation were found to exercise industry and thrift, temperance and honesty (Gordon 1981, pp. 94-114).] The crowding-in effect was substantial in total expenditure, and even the most prominent of the scientific charities became relatively generous. As the public expenditures per household were taken to the relatively low level of $3.40 per year by 1890, private expenditures rose to $15.10, ironically approaching the "lavish" public expenditures in Indianapolis, and exceeding that in Brooklyn. The crowding-in effect, measured by real payments of outdoor relief per household, was after a dozen years nearly dollar-for-dollar. Private charity, then, is expected to contribute to labor market dynamics by magnitudes not unlike that of the public relief it replaces.

The Duration Data

I drew a sample of new entrants to relief rolls from the caseworker manuscripts of the Indianapolis Charity Organization Society. The new entrants span January 1, 1881 and early April, 1881. The sample was drawn to make a strong case against the predictions of the model: real relief per household was at a minimum; labor demand was relatively strong; and the cooperation of the township trustee and of the scientific charities was at its peak. The average duration of a spell on poor relief for the 1881 entrants, then, is probably a lower bound for Indianapolis - and probably for the nation - in the early 1880s.

The sample has 102 total spells with 31 lasting longer than could be determined from the data. The average duration of a completed spell was 8 months and the average duration of all spells was 10 months. Half the spells were completed in one month. Sixteen percent of all spells lasted one year or more. The duration of spells on poor relief in Indianapolis, to repeat, was probably a lower bound in the early 1880s for the nation.

Whether the spells were short and economically important depends upon one's frame of reference. An ideal test of the hypothesis that privatization was futile would contrast the probabilities that recipients moved between the states "out of work" and "employment" before and after the sharp reduction in public outdoor relief. Such is not possible with the Indianapolis data. One of the innovations of the policy change, after all, was to introduce systematic record-keeping. Although no other micro-level studies of the late 19th-century exist for comparative purposes, one can compare the Indianapolis findings to studies across time and space (Table).



The Average Duration of a Spell on Poor Relief in the United States Has,

When Taking a Long View, Changed Little

Or: Privatization Did Not Bring Shorter Spells

Source Data Geographic Average Duration

Scope of a Spell

Hannon Yates' Sample of People Re- New York 5 mos. (New York City); (1984) ceiving Public Outdoor Relief 9 mos. (Rest of State);

in 1823; All family types; 7 mos. (New York State)

Stock Sample (All Spells)

Ziliak Charity Organization Society Indianapolis;

and Township Trustees, Lower bound

1881-1903; Oversamples for the nation

Seasonal Distress;

Flow Sample

All family types 8 mos. (Completed)

10 mos. (All)

Single-mothers 7 mos. (Completed)

13 mos. (All)

Persons Charity Organization Society, New York 1.5 years (All?)

(State of the Association for Improving the City 2 years

New York, Condition of the Poor, and 2.5 years

[1914]) the Brooklyn Bureau of Charities;

Widowed Mothers, 1914;

Stock Sample

Blank SIME/DIME, 1970-1976; Seattle and 13.3 mos. (Completed) (1989) Single-mothers on AFDC; Denver n.a. (All)

Flow Sample

Tienda Urban Family Life Survey, 1986- Chicago "less than a year"

(1990) 1987; Single-mothers on AFDC, (Completed)

1963-1987, White and Mexican; about 2 years (All)

Stock Sample

Notes: The table includes poor relief spells that were not subject to legal limits on their duration. As Hannon (1984) shows, spell lengths were for legal reasons low in New York during the 1840s and 1850s. Studies measuring spells in years are not included. Complete notes appear in the paper.

The point is clear: the average length of a spell on outdoor relief after privatization in Indianapolis was at least as high - probably higher - than that in New York before the Jacksonian reforms. And the average spells were hardly different in length from that in 1914 and from that in the Aid to Families with Dependent Children program since the 1960s (as seen from the perspective of economists nowadays, and adjusting stock samples for comparability with flow samples). The model of futility suggests there is little surprise in the constancy of spells.

A Dollar Was A Dollar

The old enthymeme on poor relief has been shown to be contradictory in theory, and the resolution of the contradiction has empirical support in a large crowding-in effect and in the constancy of spell lengths. The implicit premise in the enthymeme, of course, is that private charity is personal, and thereby more wise. Charity organizers expected that private exchanges would help create a sense of community among the richer and poorer, allowing the subtler habits of enterprise and application to rub off. The plans were not borne out by the experiment but not, as Michael Katz argued, because their "method taught dependence." In the paper I suggest the failures of personalization stemmed in part from conflicts of identity (think of Sister Carrie) and from the fact that most of the exchanges were one-shot games between the "worthy" poor and the culturally distant workers in the hybrid organizations. More importantly, personalization did not bring more and better investments in human capital.

The stories of material discontinuity between the high-tide of private outdoor relief and the emergence of the welfare state are by several measures unsound. The sometime financial and organizational efficacy of the private charities in a medium-sized city suggests that poor relief, like the lighthouse (Coase 1974 [1988]), has flourished in the absence of a large governmental provision of similar services. Despite the use of brutish imagery in prose, the city of Indianapolis did not make "worthy" applicants struggle independently as fit by nature.

On the other hand, the thesis that dependence is a welfare state problem that can be substantively reduced by privatization and by a moral reformation of a late Victorian sort, though fashionable in politics and in news editorials, is not supported by a historical referent. Nor is the thesis supported by economic theory. The local finance and administration of poor relief was dressed in scientific and moral rhetorics, but the monies entered the budget sets of recipients like other unearned monies from strangers. Sharply reducing public outdoor relief per relieved household did not raise earnings profiles above a level that compassionate people could without amelioration cope.