NOTE: The opinions and commentary expressed in this
essay are those of the author and are an exercise of free speech. They do not
necessarily represent the views of Free State Project Inc., its Directors, its
Officers, or its Participants.
by Stephen Cobb
In the universe of collective-action problems (CAPs), the Free State Project (FSP) can be categorized as a collective-migration problem (CMP). Comparing the FSP with similar problems and applying well-established theory leads to some suggestions for improving its structure and thus increasing its effectiveness. In this article we start with a brief overview of CAPs and their difficulties, and continue with a description of a special type: the CMP. Finally, we compare the current FSP structure with that suggested by theory, and recommend some changes.
We find that the FSP is subject to several classic CAP difficulties, and while it has adopted some common solutions, several remain under-utilized. The FSP remains too large, monolithic, and inflexible. The large population size has caused the FSP to suffer from both a lack of communications infrastructure and game-theoretic challenges. The FSP’s minimum participant criterion requires coordination among an unrealistically large number of people, and its one-size-fits-all Statement of Intent fails to address the varied needs of its community. Fortunately, a solution in the form of virtual community software can both facilitate communication and support an enhancement that can be implemented within the framework of the existing Statement of Intent. Enabling participants to set their own minimum commitment conditions will both better serve their individual needs and reduce the number of people requiring coordination among themselves.
1. Collective-Action Problems
1.1 Types of CAPs
When the members of a community share a goal that requires contributions from each of them, they are engaged in a collective-action problem (CAP). CAPs become interesting when it is both difficult to get contributions and difficult to exclude a member of the community from benefiting from the goal when it is achieved. Examples of these are public goods like streets that are clean, policed, and well-lit. All members of the community benefit, so we expect them all to refrain from littering and to contribute to the costs of lighting and police protection. Another example of a public good is charity: when one person makes a donation, the entire community shares in the positive feeling that the cause is being supported, so we expect that each donate according to his abilities.
An easier type of CAP is a coordination problem, where the community members do not have to contribute so much as make a common decision. For example, London clubbers might not really care which club they visit, just that it be one where enough of the right sort of clubbers are. This is actually a combination of contribution and coordination, because a good time requires a minimum number of total participants, who then must all agree on the place to go. Each clubber will have his or her own preferences, e.g. on a Monday night, when the clubs are near empty, less dedicated clubbers may prefer to go to the movies.
A third type of CAP made possible by the Internet is the mass-collaboration project, in which the contributors are distributed geographically, even globally, and may never even meet in person. Famous examples of mass-collaboration projects are open-source software like Linux, and wiki-based products like Wikipedia. While the contributors all share one overall goal, they may have slightly different visions and preferences that require compromise. They all work within the project framework in order to achieve group success.
1.2 CAP Challenges and Solutions
The more participants involved in a CAP, the more difficult it becomes to solve. This results from two factors: communications costs and free riders. The number of pairwise communications paths increases with the square of the number of participants, requiring communication and coordination infrastructure whose costs may outstrip available resources. This can be solved over time, as improving technology tends to lower communications barriers, permitting more sophisticated coordination among ever greater numbers of people distributed ever more widely. The free-rider problem appears when some members of the community find that they can reap the benefits of the project without contributing their share. The larger the community, the greater the difficulty in tracking contributions, and the less likely a community member feels a sense of ownership and obligation (a phenomenon that psychologists call "diffusion of responsibility"). Furthermore, if a community member senses that the project is more important to others than to him, he may “hold out” and try to gain more and contribute less.
Solutions to the free-rider problem include the two extremes of acceptance (that some people will not contribute their “fair” share) and coercion (usually by a government), and several more creative approaches in between. If a CAP is stuck, the following solutions may get it moving:
Form a coordination agency
CAPs may start out as being self-organizing, with each individual acting on his own and communicating directly or indirectly with other community members. For example, elderly people who prefer to live in a calm area without crime or children may slowly gravitate to a neighborhood that is close to a hospital but far from schools. As the area develops a retiree-friendly reputation, more elderly people will move there. The process would be accelerated were an entrepreneur to open a retirement home in the center of the neighborhood. The American Association for Retired People(AARP) performs a similar coordination role for pursuing the gray population’s political interests on state and national scales. Such a coordination agency thus becomes a secondary CAP that requires its own contributions. The AARP solicits membership dues (which confer some modest privileges) and then pays a staff. The AARP’s younger paid workers may not even agree with the AARP’s agenda.
If a minimum number of contributors is required to make a CAP feasible, the CAP can be approached in two or three phases. In the first phase, the coordination agency asks community members to make a public commitment to contribute, i.e. pledge. Alternatively, if the required contribution is money, contributions can be collected, to be returned if the minimum is not reached. Achieving the minimum pledges and/or contributions triggers the CAP’s implementation phase.
Increase reputational score-keeping
Contributors to the CAP tend to be compensated in one of three ways: satisfaction in the result, money, or reputation, i.e. status. If a CAP is stuck, satisfaction is presumably not enough, and soliciting money to pay workers is itself subject to the free-rider problem. Volunteers will often work for status (known as "karma" or "egoboo" in the open-source software community). This requires keeping track of who is doing what, and recognizing contributions.
Divide the problem into smaller problems
Solving a CAP is like pushing a boulder up and over a hill: a big, heavy, cube-shaped boulder will not roll easily. It may be possible to break it up into smaller, manageable pieces. If a CAP can be divided into a series of smaller CAPs, each individually justified on its own merits, they will probably be easier to solve. Ideally, the CAP could be divided into as many small goals as there are community members.
Divide the community into smaller groups
Smaller groups can focus on smaller problems, and they will suffer less from the issues facing large groups. Groups are most commonly formed on the basis of shared location and interest. Such sub-communities will tend to have stronger bonds and similar priorities. Coordination agencies thus frequently have a federated structure.
Acknowledge differential utility
In the initial description of a CAP, we assume that each member of the community is identical, e.g. shares the same preferences and priorities. This is of course not the case, and may be an oversimplification. There may be several distinct market segments, or an evenly spread continuum. Some community members will desire the goal more than others, and some community members will be able to pay the cost more easily than others. If enough community members both strongly desire the goal and can pay the cost, they may lead the effort, initiating a snowball effect.
In the technology world, the most enthusiastic market segments are known as innovators and early adopters. They are willing to pay a high price for new technology, and thanks to their demand the producers experience increasing economies of scale, leading to affordable prices for those who would otherwise not purchase the product. Figure 1 shows the classic technology adoption lifecycle.
Figure 1: Technology Adoption Lifecycle
The snowball effect can be slowed if there is a substantial gap between the first two segments and the rest. The popular book Crossing the Chasm describes the challenges posed by this phenomenon. This theory assumes that the five market segments are populated by people with specific personality types, but it seems unlikely that there would be such a discontinuity in personality distributions. The chasm effect might slow the snowball but not stop it, in the absence of other braking forces. One lesson can be drawn: the methods that worked in attracting early adopters may not appeal to subsequent segments.
An essential function of a business is to exclude non-contributors from benefiting from a collective project. Only workers are paid, and only paying customers receive a product or service. Sometimes the means to accomplish this are obvious, other times less so, bringing great rewards to the entrepreneur who can solve the problem. A related tactic is to tie a non-excluded public good to an excludable private good. For example, charities often send donors a newsletter and invite major donors to special exclusive functions. In The Logic of Collective Action: Public Goods and the Theory of Groups, Mancur Olson states that only such "selective incentives" will "stimulate a rational individual in a latent group to act in a group-oriented way." Other commonly offered selective incentives include social recognition and services such as job placement.
The boulder metaphor should be kept in mind. If the boulder is big and heavy, you can try smoothing the path, finding and motivating more pushers, or cutting the boulder into pieces that can be moved individually. The potential energy is there—its release just awaits a creative solution.
1.3 Summary of CAP Parameters
CAPs can be characterized by several parameters that describe their goal, community, and organization:
- Goal type (product, service, condition)
- Goal divisibility (monolithic all-or-nothing, divisible, or continuous)
- Goal excludability
- Goal jointness of consumption
- Community number
- Community homogeneity
- Self-organizing or actively coordinated
- Coordinators paid employees or unpaid volunteers
- Federated or centralized
- Virtual or physical
2 Collective Migration Problems
A collective migration problem (CMP) is a CAP where the community’s goal is moving to live in one territory in close association. In political migration, the ultimate goal is to achieve influence over the rules governing the territory, though other motivations are possible, e.g. economic benefit or a personal preference for association with one’s own kind. The scale can vary from the extremely local (e.g. young single people, families, or old people living in an apartment house with their peers) to the national (e.g. Jews moving to Israel). The basis of association could be religious (e.g. Mormons moving to the state of Utah), racial (e.g. the migration of black Americans northward), linguistic (Spanish-speakers moving to Miami), economic (e.g. wealthy people moving to a low-tax country like Switzerland), sexual preference (e.g. homosexuals moving to San Francisco), personal habits (e.g. non-smokers moving to a smoke-free building) or one of any number of facets of human life. The migration might be coordinated or uncoordinated based on the territory’s growing reputation. If the community values speed, time-optimal migration will require active coordination.
The most important role of the CMP coordination agency is enabling community members to signal to one another their intent. Without such an agency, community members are limited to observing each other’s actual historical movements. If the migration is over a large area, e.g. internationally, these movements may be over a long time scale and not readily apparent, and the snowball effect will be correspondingly slow. The coordination agency can enable community members to not merely announce intentions, but to bind themselves in mutual commitments. Participants can make a set of related if-then commitment rules R of the form: "If condition C is satisfied, then I will perform action A," in which many of the conditions are dependent on the rules and actions of other community members.
In a CMP, the action is generally "move within preparation time period T", though it could also be "contribute time or money to the coordination agency". The trigger condition is generally, "if N other people will do the same". At the extreme end of simplicity, the coordination agency could assume a "typical" Participant, and create a single commitment rule with some average values for T and N. Of course, individuals are unique, so every individual’s true preferences would require a large set of rules to express fully. The coordination agency must find a compromise between accuracy and simplicity, with some additional refinements to the basic commitment rule being possible.
The first and most obvious such refinement would be to let Participants choose their own minimum number of fellow movers. Participants would expect a wide range of costs and benefits, and so each would have a unique threshold at which the move would make economic sense.
A possible second refinement would recognize other types of community member: Participants (who have pledged to move if their conditions are met), Committed Participants (whose conditions have been met and are now obligated to move), and Movers (who have already moved). Obviously Movers are more valuable than Committed Participants, who are in turn more valuable than mere Participants. This relates to what economists and psychologists call “discounting the future,” and what laymen describe as “a bird in the hand is worth two in the bush”.
Each participant could be asked how many fellow Participants he or she would need to move, how many Committed participants (in the absence of other Participants), and how many Movers (in the absence of other Participants). Figure 2 illustrates the relationship among these three values. From Np, Nc, and Nm we could compute two factors kc=Np/Nc) and km=Np/Nm, which would describe how many ordinary Participants a Committed participant or Mover is worth. We would expect Np > Nc > Nm, and so km > kc > 1.
Figure 2: Participant Commitment Conditions
A third refinement would further weight or discount participants by their timing and/or likelihood of moving. A Participant with a preparation time of four years might be worth half as much as one willing to move in half the time. Participants could also have a probability attribute, based on their own self-declared likelihood of moving and/or an external assessment of their reliability.
3 The Free State Project
The Free State Project is an agreement among 20,000 activists to move to New Hampshire and exert the fullest practical effort toward the creation of a society in which the maximum role of government is the protection of life, liberty, and property. This results-oriented mission, combined with a least-common-denominator definition of libertarianism and deferment of methods and priorities, gave the FSP wide appeal and enabled it to rapidly form a large community. The FSP was established in late 2001, and by late 2003 had already attracted 5000 participants. After two years of debate, these first signers used Condorcet Voting to select New Hampshire as the future Free State. Since that time, growth has slowed from exponential to steadily linear. Many opinions and theories have been proffered to explain why.
The FSP is a combination of all three of the previously-mentioned CAP types. It is an entirely virtual organization, providing web-based infrastructure for geographically distributed participants who meet rarely if ever (this author is typing these words in a Moscow café). The FSP solved libertarian activists’ coordination problem by defining the gathering place: New Hampshire. The FSP attempts to solve the contribution problem by actively recruiting the minimum of 20,000 participants required to be successful.
And therein we find the two FSP flaws: who set the success criteria that required 20,000 participants, and is 20,000 too large a number to coordinate? The original FSP leadership based its Statement of Intent on Jason Sorens’ initial FSP announcement in July 2001 and his follow-up December article "What Can 20,000 Liberty Activists Accomplish?" These estimated that 20,000 activists (loosely defined) could attain significant influence in a state with a population of less than 1.5 million. But is that really the FSP customers’ only relevant goal, or even their most important goal? The FSP currently assumes a uniform customer base, and ignores their individual wishes, rather like Henry Ford: “You can have any color you want, as long as it is black.” The FSP’s customers have other desires as well. It is common knowledge that the FSP’s early movers are motivated by camaraderie and the increased confidence that their efforts will pay off. Some value libertarian ideals so highly that they would move anywhere, as long as the destination had a marginally higher libertarian concentration than their current home. Others work within the FSP framework and take a “free town project” approach, hoping to achieve influence with smaller numbers at the local level.
In its June-July 2006 survey of Participants and Friends, the FSP found that the most important issues that get people to sign up are:
- To show the rest of the country that libertarian principles can work
- Lower taxes
- To raise my children in a more free society
- More gun rights
Goals 2-4 can be achieved by moving to New Hampshire today--they do not require the presence of additional libertarians, though of course the more the better. Even the first goal does not require 20,000. At the beginning of 2006, with well over 100 early movers already achieving modest successes and demonstrating significant influence, it was clear that the original all-or-nothing target of 20,000 participants was a gross oversimplification. The FSP launched the First 1000 (F1K) program, a faster and leaner effort to get 1000 activists to commit by the end of the year and move by the end of 2008. This new FSP product was akin to Henry Ford introducing a second color, white, an obvious but still inadequate acknowledgment of reality. The new product sold well, and the F1K met its target, demonstrating that there really is a variety of preferences among the libertarian customer base.
This now raises the question: what other products or styles should the FSP offer? The FSP leadership moves entirely too slowly in identifying customer preferences and formulating new product offerings. This is not surprising, since the leadership also consists of unpaid volunteers with day jobs. The leadership is extremely cautious, and leery of offering an unsuccessful program. Without extensive research it is unlikely that the leadership can know what its customers want or should want. The one-size-fits-all approach ironically does not fit libertarians, who by nature prefer to make their own choices. Ideally a mechanism would be found for mass customization, to allow the customers to choose their individual products, or at least to decouple the leadership from the details of customization. Henry Ford should not be deciding the new color of the year.
The FSP organization (a coordination agency and thus a secondary CAP) suffers from a perennial free-rider problem in its lack of volunteers, but there is free-rider problem in the movement itself. FSP signups have slowed since the state vote in 2003. This is often attributed to the lack of any further exciting and motivating milestones. In the opinion of this author, the FSP reached a size at which the communications costs and free-rider effects became significant. Enthusiastic members of the FSP community frequently underestimate commitment costs (as compared with the more obvious migration costs). While the commitment trigger will occur in the discounted future, it operates over a long period of time, constraining a participant’s pre-migration life. It leads to an unrecognized cost: risk of breaking one’s promise, with subsequent reputational damage in the libertarian community. A free-rider is better off delaying his commitment until most others have done so. This was observed in the F1K program, when most participants signed up only in the last few weeks.
3.2 Way Forward
In order to serve its diverse customer base (in particular more eager movers) and overcome free-rider problems, the FSP requires a more flexible and expressive set of commitment rules without sacrificing simplicity. These commitments can fit within the existing Statement of Intent. The FSP should deploy a new registration form with expanded trigger conditions that enable participants to express more preferences:
Minimum number of Participants (maximum of 20,000)
Minimum number of Committed Participants (in the absence of other Participants)
Minimum number of Movers (in the absence of other Participants)
Preparation time (maximum of five years)
Furthermore, Participants should be able to change their trigger conditions until they have been met, perhaps with a short grace period after they have been notified. Perhaps they will even bid against each other, as those who commit to move earlier will receive more respect in the community.
The FSP leadership should remove itself from the business of divining customer preferences and devising new product offerings, and simply facilitate communication and self-organization among the FSP community members.
Figure 3 shows the first few records of a table of sample data, with participants’ trigger conditions. Figure 4 shows a table consolidating these trigger conditions (using only the Minimum Participants) into a count of participants sharing a given minimum, and then the cumulative number with that level of commitment or more. Figure 5 shows a graph of the data in Figure 4. Whenever the minimum number is exceeded, all participants with that minimum or less change status from Participant to Committed Participant.
If the trigger conditions are allowed to be in terms of Committed Participants, several iterations may be necessary to calculate Participants’ state, as each new Committed Participant may trigger more. This may be slowed by a notification requirement and grace period.
Figure 3: Example Participant Commitment Conditions
Figure 4: Example Commitment Condition Cumulative Counts
Figure 5: Graph of Example Commitment Condition Cumulative Counts
Since demand is on a normal curve, the early adopters will trigger the middle which will trigger the laggards, creating a snowball effect as shown in Figure 6.
Figure 6: FSP Community Snowball Effect
Back to Essays