Homelessness & Game Theory: How to Get the Unwilling to Do the Unwanted

Andrew Hening
8 min readNov 17, 2018

In the summer of 2018 the 9th U.S. Circuit Court of Appeals ruled that the U.S. Constitution’s prohibition against cruel and unusual punishment bars cities from prosecuting unsheltered homeless individuals for sleeping outside on public property when there is not a viable emergency shelter option. Some communities are viewing the Martin v. Boise ruling as a significant challenge to their ability to mitigate the impacts of homelessness, but when seen in a different light, the court has actually provided the impetus for a comprehensive state/national solution to the shelter crisis that is currently unfolding in cities and towns across the country.

The Impact in California

I live and work in California. While Martin v. Boise will no doubt have a national impact, it could have the biggest impact here.

Housing affordability and homelessness are at crisis levels in California, particularly in coastal areas like the Bay Area, Los Angeles, Orange County, and San Diego. Los Angeles County alone has a staggering 60,000 people living on its streets every day. There are another 20,000+ people in the Bay Area. Despite being 12% of the nation’s population, California has 25% of the country’s homeless population.

Making matters worse, California only shelters 32% of the homeless community. This means California has the highest unsheltered homeless rate in the entire country.

The Systemic Dilemma

Homelessness is a symptom. It is the result of a breakdown in other upstream systems — housing, employment, addiction, mental health, systemic racism.

These systems are fairly equitably distributed throughout the country. However, to the extent to which higher housing prices are positively correlated with higher rates of homelessness, that would explain why increasingly expensive urban housing markets are the only parts of the country witnessing increases in homelessness.

Despite any professed goals or policy dictates from the federal government, there is not currently a national, federally funded framework for comprehensively preventing and ending homelessness.

As a consequence, for all intents and purposes, the roughly 22,000 city, town, and county governments across the United States are each independently responsible for addressing homelessness in their own local way.

If you are a local government, on the extremes, you essentially have two options for addressing homelessness. You can either Option A. Attempt to house and provide services for everyone who is homeless or Option B. Attempt to relocate all homeless people out of your community. There are three primary reasons local governments would NOT want to pursue Option A:

#1 The cost of providing housing and services is substantial (especially without federal support).

#2 There is a pervasive belief in every community that the majority of people who are homeless are NOT local. This belief signals an assumption that other cities have ALREADY selected Option B. Unfortunately, this “disbelief” is contrary to overwhelming empirical data. 70 to 85% of people experiencing homelessness reside in the same county where they were previously housed.

#3 Concern about this second point results in a first mover “disadvantage.” Since the upstream causes of homelessness are happening in every community, unless all communities move in parallel to offer housing and services, then people who need help (and cities looking for a resource) will begin to gravitate towards first movers.

This struggle between Option A and Option B is analogous to the well known game theory problem the “Prisoner’s Dilemma.”

The Prisoner’s Dilemma analyzes the decision matrix of two criminal accomplices who are interrogated separately by the police. Each criminal can either keep quiet or testify against the accomplice. If they both keep quiet, they will both receive a minor sentence. If they both testify against each other, they will both receive more severe sentences. If only one person testifies against the other, the confessor will go free and the other person will get the most severe sentence. As an individual, the best option is to testify against the accomplice and thus go completely free. However, since both individuals have this same incentive, if both individuals pursue their best interests, they will both get more severe sentences. Therefore, the optimal solution is for both individuals to commit to cooperate by remaining silent. They will both suffer slightly in this scenario, but they will avoid the longest sentences.

The Prisoner’s Dilemma exemplifies the choices faced by individual actors within a “Tragedy of the Commons.” Tragedy of the Commons is a term used in social science to describe a situation in a shared-resource system where individual users acting independently according to their own self-interest behave contrary to the common good of all users by depleting or spoiling that resource through their individual actions. Examples include:

  • Overcrowding — People love national parks because of their beauty, wilderness, and solitude. Because people love the parks, more and more people try to visit the parks, creating more traffic, litter, and noise. The parks lose some of their beauty, wilderness, and solitude because of the overcrowding.
  • Air Pollution — Imagine a bunch of people shopping for cars. They could either buy cheap cars that emit a lot of dirty exhaust, or they could buy more expensive cars that emit less exhaust. Individually, everyone would be better off financially by buying the cheaper, dirtier cars, but in the long-run everyone is hurt — physically and financially — by bad air quality.
  • Trash — Plastic was one of the most important inventions of the 20th Century. Today it provides extraordinary convenience our food and beverages. Unfortunately, by consuming what is easy and convenient, 40% of all plastic is now used only once and then discarded. Less than 10% is recycled. 18 billion pounds of plastic waste flow into the oceans every year.
  • Society’s response to homelessness has become a Tragedy of the Commons. It is in each community’s individual best interest to NOT try to solve homelessness, but in the process, homelessness overall continues to get worse and worse to the detriment of all communities.

The Solution

Fortunately, there are a variety of strategies for effectively addressing a Tragedy of the Commons:

1. Educate people about the damage that is being done.

2. Use moral pressure to encourage self-policing.

3. Internalize externalities by finding ways to make people feel or be responsible for the negative consequences. For example, a gas tax factors in the societal cost of building roads and addressing air pollution.

4. Privatize the common resource, so everyone becomes individually responsible for a smaller piece.

5. Create government regulation that stipulates how much of the resource can be used. For example, people and companies might have to compete for a limited number of permits, or there might be a quota system on a given resource.

6. Put a price on the common resource. Even in water-scarce areas like the American West, water is largely free for the end user. If there was a price on water after a certain guaranteed free amount, wasting water on big yards or other activities might not make economic sense and would thus lead towards better conservation.

Historically advocates have relied on education and moral pressure to compel action around housing and shelter. These strategies have clearly been insufficient.

Given this historical failure, Martin v. Boise is a major milestone. It has finally moved society beyond just education and moral pressure. The ruling has internalized externalities by providing people experiencing homelessness with a legal recourse to punish (i.e. sue) cities that resort to Option B strategies.

The challenge, however, is that there has historically been an UNEQUITABLE distribution of Option A and Option B responses across cities and towns. Thus, this ruling, while well-intentioned, has the potential to put even more pressure on “first mover” cities and towns that have already invested in providing housing and resources. While there is a very high degree of “localness” at the county-level, that doesn’t necessarily hold for individual cities and towns within a county. Since significant numbers of people experiencing homelessness have already gravitated regionally to these cities and towns, these communities will continue to experience disproportionate regional inflows while now having fewer options for mitigating negative impacts.

Thus, for first mover cities and towns, the question now is — is there a way to leverage Martin v. Boise to ensure that ALL communities pursue Option A responses to homelessness? The short answer is “yes.” This can be done by leveraging the final three strategies for addressing Tragedy of the Commons. For this I’ll use San Francisco as an example:

#1 As a first mover, San Francisco should “privatize” homelessness by guaranteeing a right to shelter for any person that can show their last verified address was in San Francisco. For any non-San Francisco homeless residents, Martin v. Boise should not apply, and cities should still be able to enforce against camping on the streets.

#2 Next, San Francisco should build a coalition of other first mover communities to lobby the State of California to create a “regulatory environment” wherein homelessness is “privatized” in every community. This would be done by legislating that any person who becomes homeless in California has a right to shelter in the city/town/unincorporated area where they last resided.

#3 In exchange for this “stick”, the State should then “monetize” the problem by simultaneously guaranteeing additional funding to local communities to increase supportive housing services.


The past 40 years have witnessed a complete and utter societal failure to comprehensively address homelessness. By recognizing the underlying systemic dynamic that has stymied action, we can finally start to strategize a new approach.

The reality is that the current scarcity of shelter / housing / services is already a refusal to residents / non-residents alike. By creating a regulatory environment that requires every community to make a small “sacrifice” through creating some guaranteed amount of shelter / housing / services, we will finally have a shared framework for cooperation that will begin to address this issue at scale.

Importantly, by creating a feedback loop that requires action on the part of local jurisdictions, there will be a greater incentive than ever to prevent homelessness in the first place. The desire to avoid creating shelter and other programs might finally spur the necessary political and financial will to address the upstream causes of homelessness — housing affordability, labor participation, living wages, recovery services, mental health services, systemic racism.



Andrew Hening

UC Berkeley MBA and Harvard-recognized culture change leader sharing tools, strategies, and frameworks for untangling complex and messy challenges.