Professional Documents
Culture Documents
Abstract
JEL classijication: L3 1
1. Introduction
Entrepreneurs play a central role in proprietary firms. They are the individuals who
decide to found a firm, generally in return for the right to appropriate the profits
generated by its activities. The simplest theoretical model of their behavior assumes that
they will seek to maximize their firm’s profits. Nonprofit firms are a different sort of
organization, whose numbers have recently been expanding more rapidly than proprietary
firms’. These firms can be profit-making but are legally forbidden to distribute any
surplus to their founders. Yet there are individuals who devote their entrepreneurial
talents to founding and operating such firms. The question is: why? The economic
literature on these enterprises yet does not contain a commonly accepted model of the
behavior of nonprofit entrepreneurs. This paper develops such a model.
The aspect of nonprofit entrepreneurial behavior we are interested in explaining is the
following: Which member of a heterogeneous population will be the one to found and
operate a nonprofit firm ? Specifically, what characteristics should we expect such an
individual to have?
Previous work directed at this question has been primarily in the form of statistical and
case studies3. In a comparative study of nonprofit organizations around the world, James
(1987) concludes that “... universally, religious groups are the major founders of nonprofit
service institutions”. Young (1985) reports on a series of case studies which he
characterizes as revealing a wide variety of motivations and backgrounds among
American nonprofit entrepreneurs. Young (1983) proposes a set of stereotypes meant to
capture alternative motivations and entrepreneurial styles. No formal or predictive
behavioral model, however, is presented in this work. One recent contribution containing
a theoretical model can be found in (Eckel and Steinberg, 1993). Their model frames
entrepreneurship decisions in terms of career choices by individuals who could get a flow
of pecuniary and nonpecuniary incomes in either the for-profit or nonprofit sector. They
assume that the nondistribution constraint on the net cash flows of nonprofit enterprises
only constrains the form (not the amount) of profit distribution, so that nonprofit
managers can capture some of the firm’s revenues in the form of managerial perks. They
then show that potential entrepreneurs include both perk lovers and public good lovers
(individuals who care about the quality and quantity of the firm’s output).
The type of enterprise we are most concerned about in this paper is a philanthropic
organization, i.e. a nonprofit firm which provides a public good and derives at least some
of its revenues from private voluntary donations. To model the choice to found such an
organization, we assume that individuals get utility from both their private consumption
and the public good, and so are willing to contribute some of their private resources
toward having the public good provided. However, private provision cannot occur unless
someone (the nonprofit entrepreneur) accepts the task of organizing a firm to collect
private donations and put them to use in this way.
We assume that the individual who volunteers his entrepreneurial services incurs a
private cost from doing so before any of the public good is provided. These costs could
include the direct expenses incurred in setting up the organization, advertising its goals,
and executing the fund-raising activities, as well as the opportunity cost of the time he
spends in these activities. After this, the rest of the population can contribute to the firm,
which then uses these resources to provide the public good. Incurring these sunk costs
‘Between 1967 and 1983, the number of nonprofit firms operating in the U.S. grew by 170%, whereas the
number of proprietary firms increased by less than 100% over the same period (see Weisbrod (1988), Table A. 11
p. 169).
3James (1987) and James and Rose-Ackerman (1986) provide partial surveys of this work.
M. Bilodean, A. Slivinski/J. of Economic Behavior & Org. 31 (1996) 117-127 119
may be a profitable investment for the entrepreneur if it allows him to capture a flow of
managerial perks, get a warm glow feeling, enhance his career opportunities, or use his
position to influence the characteristics of the public good to be provided. Then, the
nonprofit entrepreneur would simply be found in that subset of individuals for whom
nonprofit entrepreneurship is privately profitable.
However, it is also quite possible that the entrepreneur would have to work long hours
and accept a low pay just to keep the organization afloat, and that perks would be few and
far between. When private rewards are too small to completely offset the start-up costs,
who will make a sacrifice for the common good? If nonprofit entrepreneurship is
privately costly for everyone, then they would prefer to let someone else undertake it.
Simply waiting for this to occur, however, prolongs the delay until the utility stream from
the public good can begin. This trade-off between privately incurred costs and public
benefits is also analyzed in Bliss and Nalebuff (1984) Bilodeau and Slivinski (1996).
Here, it results in a ‘war of attrition’ game being played by members of the population.
That is, a game in which time passes continuously, with no public good provided, until
someone ‘volunteers’ to found a nonprofit firm. That individual then incurs a private cost
for some period of time, after which others may contribute some of their private wealth to
the provision of the public good, and finally the utility flow from it accrues to all
members of the population.
For any individual i in this game, we can construct a parameter ti, which is the point in
time beyond which it is no longer worthwhile for i to volunteer. We then show that there
is a unique subgame perfect equilibrium in which the individual with the largest fi value
volunteers at time 0 to take on the role of nonprofit entrepreneur. The value of ti is shown
to depend on:
The dependence of the value of ti on these parameters then allows for comparative
statics results which generate predictions regarding the characteristics of those
individuals who volunteer. Not surprisingly, the model predicts that individuals whose
cost of entrepreneurship is low, possibly because they have superior organizational skills
and/or prior experience in similar activities, or who get a large increment to their utility
stream from the public good, possibly because they are highly sensitized to the need for
the good in question, are more likely to end up in this role. The model also generates a
number of less obvious predictions. It predicts that nonprofit entrepreneurs will tend to be
relatively impatient individuals in the sense that they have a high marginal rate of time
preference. At the same time, the model predicts that nonprofit entrepreneurs will also
tend to have a relatively long time horizon. That is, either they are relatively young, or
care about the benefits that would be generated by the firm for individuals who will be
alive far into the future. Also, the model predicts that if the nonprofit firm will produce a
120 M. B&dean, A. Slivinski/J. of Economic Behavior & Org. 31 (1996) 117-127
good which will induce high levels of contributions from the public, then the entrepreneur
will tend to be a relatively wealthy individual; someone from the upper tail of the wealth
distribution. Alternatively, if private contributions to the fii tend to be small, then the
founder is more likely to come from the middle of the wealth distribution.
2. Model
‘Individuals whose wealth is too small to ever volunteer may still contribute to the provision of public services
if someone else volunteers, of course. Including such individuals in the analysis is trivial, and changes none of
the results.
sWe cannot rule out a priori the possibility that two or more individuals may volunteer simultaneously to find
rival nonprofit entreprises competing for donations. We therefore define Ui/ as i’s stream of payoffs if subset J of
individuals simultaneously found nonprofit entreprises. In this case, i could contribute to all, some or none of
these firms, and Z” would depend on how each entrepreneur allocates in equilibrium the donations he receives
between the various characteristics of the public good. For our purposes here, it is sufficient that Z” simply be
well-defined for any J, so we assume that this is the case.
M. Bilodean, A. Slivinski/J. of Economic Behavior & Org. 31 (1996) 117-127 121
3. Analysis
is i’s payoff if the set of individuals in J all simultaneously found nonprofit firms at time I,
where .I is any subset of individuals not including i. Further, it follows that
is the payoff to i if the public good is never provided during his time horizon. We assume
that Z_+(O)> Si(Ti)V i so that despite the entrepreneurial costs, everyone would rather
volunteer immediately than never consume any public good. Note that since
then Li(Ti ~ A) < Si(7;), and therefore, Z+(t) < Si(Ti) for all t E [Z’i- A, 7;[. It follows
that there is, for each i, a unique time ti beyond which it is no longer worth volunteering
to be the entrepreneur who organizes the provision of public good, even if this means that
the good will not be provided during i’s horizon. There are two broad possibilities to be
considered:
1. FiJ(t) > L,(t) for all t,i,J. This is the situation that will arise if the private rewards to
nonprofit entrepreneurship (perks, warm glow, enhanced career opportunities, etc....)
are small or not important to i, and if characteristics of the public good to be provided
would not vary much across entrepreneurs, relative to the cost of entrepreneurship.
Then volunteering nonprofit entrepreneurial services is privately costly and everyone
would rather let someone else do it.
2. FjJ(t) < &(t) for some i,J. This will occur if the entrepreneurship costs are low, if
there are sufficient perks and other nonpecuniary rewards to the nonprofit entrepreneur
or if the characteristics of the public good provided vary sufficiently across potential
entrepreneurs. Then despite the initial cost of entrepreneurship, this is a desirable job
and i would rather do it than let some other subset J of individuals take charge of
providing the public good.
6Note that when individuals have different preferences for the characteristics of the public good, it is possible
that FtJ(r) < &(r) even when i has to pay at least a part of the entrepreneurial costs, because he might be able to
allocate his donation and those he collects from others (if any) in a way that improves the equilibrium
characteristics mix. We leave this issue aside here, and simply assume that F,J(~) > &(t) Vt, Vi, J.
M. Bilodean, A. Slivinski/J. of Economic Behavior & Org. 31 (1996) 117-127 123
1
Uii - Vi
li = Ti - Xln (Uii _ yi)e-‘;a _ (& _ vi)
[
Note that since uii > vi, the assumption that Li(0) > Si(7;:) implies that the
denominator of the expression in brackets is positive, so that A>0 implies that the
expression in brackets must be greater than 1. Hence ti E [0, Z’i- A] for every i. Since the
individual with the largest value of ti will volunteer immediately in equilibrium, the
interesting questions concern the factors that influence the value of ti.
It is immediate that ti is increasing in T, so that ceteris paribus the nonprofit
entrepreneur will be the individual who would stand to benefit from the public good for
the longest time. In particular, this means that nonprofit entrepreneurs will tend to be
relatively young, or, if the horizon Ti is interpreted to potentially include the lifetimes of
future generations (SO that Ui is then a ‘dynastic’ utility function), then those whose
concern for future generations extends furthest into the future will tend to take on this
entrepreneurial role. A simple calculation also shows that ti is increasing in ri, the
individual’s marginal rate of time preference. So nonprofit entrepreneurs will also tend to
be impatient individuals who discount their future at a high rate.
Suppose now that individuals are identical except that they would incur different costs
while setting up the firm. Thus, only ci varies with i. Then
dti
-_=
dUi/dCj[l - e-q <o
act ri[(4i - yi)e-“* - (VI - y,)]
Not surprisingly, the individual who would incur the smallest entrepreneurship cost,
possibly because he is more skilled at this type of activity, or has prior experience in a
similar enterprise, will be the entrepreneur.
An interesting question with a less obvious answer has to do with the effect of wealth
on the propensity to volunteer. To investigate this, assume now that individuals are
identical in every respect other than their flow of income, Wi. Now consider a small
redistribution of income among individuals and toward i which leaves the aggregate
wealth of contributors unchanged. Then we have:
where
1 _ c-r,*
A = Ti(Uii - Vi)[(Uii - Vi)fY’lA - (Vi - yi)]
The assumption that L+(O) > Si(Z’i) implies that A>O, SO the sign of &i/&i depends on
the expression in brackets and is ambiguous in general.
To sharpen these results, assume now that preferences are separable, in that
Ui(Xi, Z) = g(Z) + h(G), f or all i, with both g(.) and h(.) increasing and concave. Then
Eq. (1) becomes:
This is positive since h(.) is concave, so among noncontributors ti increases with wi.
Intuitively, with separable preferences all noncontributors get the same benefit from the
M. Bilodean, A. Slivinski/.l. of Economic Behavior & Org. 31 (1996) 117-127 125
public good, but if there is decreasing marginal utility of private consumption, the cost of
entrepreneurship weighs less heavily on wealthier individuals. Thus, poorer noncon-
tributors are less likely to become nonprofit entrepreneurs, simply because they can least
afford to bear the costs of entrepreneurship.
Suppose now that i is a contributor. If her income is increased via a transfer from
another contributor, then Warr’s theorem (Warr, 1983) implies that wi - zT is unchanged
for all contributors, because everyone will just change their contribution levels by the
amount of the change in their income. For a small change of this sort, then, it follows that
h’(wi - z;) = 0, so that Eq. (2) reduces to:
The sign of this expression is ambiguous if h(.) is concave, with the first term in
brackets being positive and the last negative. Intuitively, everyone gets to consume the
same quantity of the public good, but when preferences are identical, every contributor
also ends up consuming the same quantity of private goods. Richer individuals just
contribute more to the public good, so that in equilibrium all contributors have the same
level of utility once the nonprofit firm begins operation. However, until the public good is
provided, poorer contributors have a lower level of well-being. Therefore, poorer
contributors have more to gain from the provision of public good and this tends to make
them more willing to volunteer. On the other hand, when there is decreasing marginal
utility of private consumption, poorer contributors can also less easily absorb the costs of
entrepreneurship and this tends to make them more reluctant to volunteer.
The sign of Eq. (3) hinges on the degree of concavity of h(.). In particular, if h(.) is
linear, implying constant marginal utility of private consumption, the first term vanishes
and Eq. (3) is negative. As concavity of h(.) increases, the first term grows and eventually
comes to dominate, making the expression positive.
To provide an interpretation of the reason why the concavity of h(.) influences how
income affects the propensity to become an entrepreneur, consider a population of
individuals with identical preferences of the form U(xi, z) = h(xi) + g(z), and let Z’ be
the equilibrium level of contributions to the nonprofit firm in this population. (With
identical preferences, this will not vary with the identity of the entrepreneur.) Now,
consider that same population, now sharing the alternative utility function
V(Xi, Z) = k(xi) + g(Z)7 and let Z2 designate the resulting level of contributions. Further,
assume that k(.) is ‘more concave’ than h(.), in that h’(0) = K(O), and for every
x 2 0, 0 > h”(X) > k”(X).
Then one can show that it must be that Z2>Z1, and that individuals in the popu-
lation with k(.) contribute more from the same wealth. Thus, the identical population
contributes more to the nonprofit firm when it collectively has a more concave sub-utility
function for private consumption. Thus, it follows that in populations in which individual
preferences lead to relatively high levels of individual and total contributions, it will
tend to be the wealthier contributors who take on the entrepreneurial role, since ti
then increases with Wi. In situations in which contributions are smaller, ti will increase
126 M. Bilodean, A. Slivinski/J. of Economic Behavior & Org. 31 (1996) 117-127
t,
c E
with w, among the noncontributors, but decrease among contributors, and the
entrepreneur will tend to come from the wealthiest of that set of individuals who
contribute little or nothing.
Fig. 1 plots ti against wi for a population in which only wi varies across i, and
preferences are separable, with h(.) not being ‘too concave’, fi increases from c to J (the
income of the individual who is just indifferent between contributing or not) and
decreases thereafter. Since the volunteer will be the individual with the largest value of ti,
the model predicts that ceteris paribus, the volunteer will be the wealthiest noncontributor
(i.e. the wealthiest among the n - m poorest individuals). The nonprofit entrepreneur
would in this case be a ‘middle class’ individual who would not have contributed
voluntarily to the provision of public good if someone else had solicited him, but who is
willing to invest time and money to solicit donations from those richer than he. A very
concave h(.) would cause ti to increase with Wi even among wealthier individuals, in
which case the richest individual in the community would be the one most willing to
volunteer nonprofit entrepreneurial services.
Maintaining the separability assumption, we can also consider a population in which
only the function gi(.) differs across individuals. Note that this function enters into the
definition of ti only via the term Gi = gi(Z) - gi(O), and that ti is increasing in Gi. In such
a population then, it would be that individual who gets the largest personal payoff from
the public good who would found the nonprofit firm. This is the closest analog in our
model to the private incentive to set up a proprietary firm.
M. Bilodean. A. Slivinski/.l. of Economic Behavior & Org. 31 (1996) 117-127 127
4. Conclusion
We have examined how various factors could influence the decision to supply
nonprofit entrepreneurial services to organize the provision of a public good. The model
proposed emphasizes the desire to consume the public good itself as the entrepreneur’s
primary motivation for doing so, but admits a wide range of other motives such as
managerial perks, career concerns, feelings of warm glow and seeking to impose one’s
views. The analysis presented focusses on the case where nonprofit entrepreneurship is
privately costly so that everyone would prefer to let someone else do the job and makes a
number of predictions regarding the characteristics one should see exhibited by the
nonprofit entrepreneurs who found and operate nonprofit firms.
It would be most useful to see if these predictions are borne out in actual data,
recognizing of course, that the model presented here does not capture every aspect of
nonprofit entrepreneurial motivations or behavior. In particular, uncertainty and attitudes
toward risk are central factors in the self-selection of nonprofit entrepreneurs that are not
considered here. The data needed to confirm these predictions empirically are
information about the age, income, and other visible characteristics of nonprofit
entrepreneurs, as well as about the entrepreneurial act itself. Founding a new firm is an
identifiable entrepreneurial act, but so is setting up a new program within an existing
organization or expanding an organization’s reach into new geographical areas. The
entrepreneurial effort is itself also an ongoing process, sometime involving a fluid group
of individuals to various degrees over a number of years before coming to fruition. A
good starting point for the required data can be found in the case studies of Young (1985).
References
Bilodeau, Marc and Al Slivinski, 1995, Rival nonprofit firms, Working paper, Universite de Sherbrooke.
Bilodeau, Marc and Al Slivinski, 1996, Toilet cleaning and department chairing: Volunteering a public service,
Journal of Public Economics 59, 299-308.
Bliss, Christopher and Barry Nalebuff, 1984, Dragon-slaying and ballroom dancing: The private supply of a
public good, Journal of Public Economics, 25, I-12.
Eckel, Catherine and Richard Steinberg, 1983, Tax policy and the objectives of nonprofit organizations, Working
Paper, Indiana University Purdue University, Indianapolis.
Fudenberg, Drew and Jean Tirole, 1991, Game Theory, (MIT Press, Cambridge, MA).
Hendricks, KenWeiss Andrew and Wilson Charles, 1988, The war of attrition in continuous time with complete
information, International Economic Review, 29, 663680.
James, Estelle, 1987, The nonprofit sector in comparative perspective, In: W.W. Powell, Ed., The Nonprofit
Sector: A Research Handbook, (Yale University Press, New Haven, CT) pp. 397-415.
James, Estelle, and Susan Rose-Ackerman, 1986, The Nonprofit Enterprise in Market Economies, (Harwood,
New York).
Warr, Peter G., 1983, The private provision of a public good is independent of the distribution of income,
Economics Letters, 13, 201-207.
Weisbrod, Burton, 1988, The Nonprofit Economy, (Harvard University Press, Cambridge, MA).
Young, Dennis R., 1983, If Not for Profit, for What? (Lexington Books, Lexington, MA).
Young, Dennis R., 1985. Casebook of Management for Non-Profit Organizations, (Haworth Press).