The Pitch, Part Two: Return on Investment

As promised in last week’s blog, I will focus this week on the return on investment (RoI) of The Change Campaign’s work. The Full Press is the work that has a quantifiable and financial RoI, so that is what I will focus on in this blog post. As I have shared before, our society’s lack of wellbeing costs us a tremendous amount of money, in the trillions of dollars every year. A 2023 JAMA article estimated that the annual cost of Adverse Childhood Experiences (ACEs) in the United States is $14.1 Trillion. At the same time we have seen Washington State implement a program that reduced ACE prevalence at a population level (which has come to be known as Self Healing Communities and ran from 1994-2011) and saved the state over $1 billion. ACEs are a powerful population-level prediction tool across physical, psychological, and social health outcomes. Population attributable risk (PAR) data from various sources shows ACEs contribute to: 25.5% of cardiovascular disease; 24.3% of cancer; 22.2% of asthma; 55.7% of anxiety; 42.7% of medical treatment for a mental health condition; 58% of lifetime suicide attempts; 28% of lifetime smoking tobacco; 65% of alcoholism; 54% of using painkillers to get high; and 78% of intravenous drug use. These are just some of the known PAR values associated with ACEs, but I share them to show the wide range of associated outcomes and subsequent costs. On top of these models, Maryland developed a prevention-oriented credit-based methodology to save the state millions of dollars by promoting better outcomes (this model is specific to diabetes, but we can build out across multiple ACE-attributable outcomes). Our work is about preventing more than just ACEs, but because they are finite they are quantifiable, which for the purposes of this piece and creating an economic model, ACEs are a valuable tool. Leveraging the economic model proposed in The Cost of Doing Nothing framework, we can take these models and create a real estimate of what The Full Press’ RoI could be. These are far from all the sources that have gone into creating a framework to estimate and measure the RoI on this work. I will spare the length it would take to describe it all, these were picked to highlight the economic reality and proof of concept, but if you would like to see more, my capstone goes into greater detail. 

That was a beefy paragraph, but before I get into the pitch side of the RoI of The Full Press, I wanted to establish that this is built on sound reasoning. This is not some fantasy, but is based in the reality of what we know we are dealing with, and what we have seen work. If we were to reduce the current trajectory of ACE prevalence by 0.1 in the state of Pennsylvania (according to Behavioral Risk Factor Surveillance System data), which mind you would actually still be a rise in ACE prevalence because of current trends, we would save the state over $280 billion in lifetime costs across current generations. What is even more profound is the compounding interest of this RoI across generations as we prevent ACEs. As children are born into families with healthier legacies that block rather than perpetuate the intergenerational transmission of trauma, we would see exponential RoI for multiple generations. As with any exponential curve there would be a plateau phase, but we could expect that to be many generations in the future, and at that point the incentives of systems would be very different than they are today. 

The RoI I am proposing is massive. We will never prevent all ACEs (nor trauma broadly), there are bad things that happen that impact us deeply, which is one of the prices we pay to live this beautiful life, but we can prevent a lot of ACEs and trauma that are experienced. So I will not claim the RoI to be $14.1 Trillion per year, as that would be disingenuous. As I will criticize later, it seems to be normal for people pitching on their potential RoI to claim every theoretically possible dollar as a potential return, but I want to be genuine in what I believe is possible. To look at estimates of what minimum investment is needed and what returns might look like, we will use both the state of Pennsylvania and the entire United States for examples. 

First, the assumptions that underlie the costs. We are assuming that we are funding county coalitions at an inflation-adjusted level. The aforementioned Washington State model spent about $80,000 per community per year, which in today’s dollars is about $150,000. The state of Pennsylvania has 67 counties, which would mean an annual investment of $10,050,000. The model also assumes a coordinating entity for the state with a budget of $2 million. So the full annual cost to the state is assumed to be $12,050,000. Looking at the costs for the entire United States, we assume the same county level of cost per county across the 3,143 counties in the country, the full costs of which would be an annual investment of about $470 million. We would then multiply the statewide coordinating entity cost of $2 million across all 50 states, adding another $100 million, plus an additional $20 million national coordinating entity, so the total annual US cost is estimated to be $590 million. Now, as with anything that grows, we must not assume that we can come in and create something out of nothing all at once, so we do not have those full annualized assumed costs each year in the model. In year one, spending is 10% of its full amount, and taking 10 years to reach full capacity at a linear rate. We are measuring the RoI over the course of a generation, so running the model over a 20 year period, the ultimate cost to the model is $186,775,000 in Pennsylvania and $9,145,000,000 in the United States. 

Again, these are not definite figures, but assumptions that enable us to test a model. Now that the costs are estimated, what about the returns? Rather than assuming that we would achieve the same results as Washington did, though with the increased knowledge and technology available to us since that model ran I think that we could not just meet but exceed the results they realized, the model is run with a low-end return of 3:1 (less than 10% as effective as Washington), a mid-level return of 10:1 (less than a third as effective as Washington), and a 35:1 (the RoI measured in Washington’s own calculations). We assume in the model that there is no return in the first six years, and then the returns begin to ramp up, based on the costs from six years prior multiplied by the aforementioned low, medium, and high multiples. By this logic, it is not until year 16 in the model that the returns hit their full amount. So in year one, the cost for Pennsylvania is $1,205,000 and the cost for the United States is $59,000,000. The returns for both in that year are assumed to be $0. The first year that returns emerge in the model is in year seven. At this point the cost for Pennsylvania is $8,435,000, and the returns are modeled at: low end $3,615,000; middle tier $12,050,000; and high end $42,175,000. The cost for the United States in year seven is $413,000,000, and the returns are modeled at: low end $177,000,000; middle tier $590,000,000; and high end $2,065,000,000. The other assumption built into the model is that only a quarter of the returns are actually capturable, rather than returns by some benefit that is not able to be monetized. All of this aims to create a reasonable model.

With these assumptions, after 20 years, Pennsylvania would realize a capturable RoI (again, taking a quarter of the total returns) of $86 million on the low end (0.46 returns per dollar spent), $286 million at the middle tier (1.53 returns per dollar spent), and $1.00 billion at the high end (5.36 returns per dollar spent). The United States running the same assumptions would realize a capturable RoI of $4.2 billion on the low end, $14.0 billion at the middle tier, and $49.0 billion at the high end. The returns per dollar spent ratios are the same in the United States as in Pennsylvania at each level because the model is the same. So, with this projection, the low end is unsustainable and the break even point for capturable RoI to governments is a 6.53:1 total RoI, which is less than 20% as effective as what Washington State was able to realize. There is reason to believe that we can exceed the break even point and make this sustainable. 

Now, I tried to make good faith assumptions, but there are limitations in this model. For instance, I didn’t include a 3% discount rate on the returns because I didn’t want to get more bogged down in trying to explain the model than I already was, but if I had, then the break-even would rise to about 7.1:1, which is still well within reason. Similarly, we also kept the costs consistent from year to year outside of the factored in ramp up period, which is unrealistic, but models general truths in a defensible way. There is great likelihood that as the work scales we would also invest at a sub-county level, but doing so again became more complicated than I felt was useful for the purposes of this blog post. Additionally, the model stops at 20 years and holds returns flat once they reach full strength, so it does not capture the compounding across generations that would likely show up after that, which likely short changes the returns we could responsibly model. I will continue working to improve modeling to capture full defensible costs, yet, even given the limitations discussed here, this model turns out to be incredibly hopeful.

Some people who read this may be thinking: This is way too much money! Have you lost your mind? The answer to that question is maybe, I am biased and definitely the wrong person to answer the latter question. But to the former statement, I encourage you to look at the current investments and losses in AI. Recent reports tell us that Anthropic had $8 billion in operating losses and net losses of $42 billion in 2025 building something that their CEO claims has a decent chance of destroying humanity. We are building something that even if zero returns are realized, it would take us a long time to reach that level of spending, and on the other end of it we would have helped communities and individuals thrive. I am confident that we will realize more than zero returns, but still, I believe in the value of what we are building beyond the RoI. 

There is so much investment going into smaller categories of spending that have questionable returns, creating concentration risks. There is concern that the AI build out threatens a market correction, and meanwhile in the government we have seen our budgets move more money into militarization domestically and abroad, and simultaneously cut spending in health, education, and other services. I won’t make a moral argument about this as I close this out; it would take this blog into a completely different topic, but I know the money for an investment like this exists. 

I am tired of being told that this space is just not investable while other unproven markets receive unbelievable influxes of capital that are destabilizing people’s lives and our economy. Until there is real investment into The Full Press I cannot be positive that it will work, but there is every reason to believe that it can and will, and I will not stop trying to not only fund the work, but more importantly build the capacity that helps people and communities live better lives. This blog is another effort at taking a step in that direction, and I hope it is the step that gets us there, because I know so many in the United States and around the world want something like this.