Can A $239 Million Prison Transformation Pay for Itself? 

California’s $239 million investment in the San Quentin Rehabilitation Center has become one of the most debated criminal justice expenditures in the country. When announcing the transformation, Governor Gavin Newsom framed the project in practical terms: ninety-five percent of people in prison will return to communities, and the question is what kind of neighbors they will become. 

The redesign of San Quentin expands education, workforce training, and structured reentry planning inside the facility. Critics see the renovation as an expensive gamble. Supporters view it as overdue modernization. What has largely been missing from the debate is a structural analysis of what happens financially if rehabilitation outcomes change.

To address that gap, The Last Mile worked with researchers to build a multi-year structural model tracking how evolving recidivism and employment outcomes under the California model influences overall return on that investment. The model compares a status quo trajectory with a rehabilitation-driven trajectory and calculates how those differences affect incarceration costs, lost tax revenue, productivity, and cumulative savings. Instead of evaluating a single cohort, the model follows annual release groups from 2026 through 2035, allowing changes to compound. The central question becomes whether sustained improvements across a decade can offset the initial investment.

A Structural Shift: Why More People Will Be Released Through San Quentin

The model begins with a structural assumption that more people will be released through San Quentin over time. 

Rather than operating as a traditional prison, the function of the redesigned facility will shift into a pre-release hub where individuals spend their final years preparing for reentry. Under this assumption, annual releases increase from just over 250 individuals in 2026 to roughly 900 by 2035. 

This change does not assume more people are released statewide. It assumes individuals nearing release are routed through San Quentin so they receive education, employment preparation, and reentry planning before returning home. As more people move through the facility, more individuals are exposed to programming designed to reduce recidivism and increase employment.

Projected Releases Per year from San Quentin

The True Cost of Recidivism Extends Beyond Prison Budgets

Recidivism carries costs beyond incarceration alone. The model includes multiple components to capture the full impact, including pretrial detention, prosecution and courts, public defense, victim services, parole supervision and more.

Each return to prison removes a worker from the economy and triggers a new list of public expenditures. Under the status quo system, these costs repeat consistently. Under the new rehabilitation model, fewer people return over time, reducing both direct and indirect costs.

The model calculates a comprehensive cost of recidivism. Direct incarceration costs are estimated using average annual cost multiplied by an assumed 1.3-year return length. The model also includes broader social costs (like those listed above), and lost tax revenue is calculated using projected wages and effective tax rates. Productivity losses account for income that individuals would have earned if they remained in the workforce. Together, these components create a full cost of recidivism that extends far beyond prison budgets.

These costs are calculated for both the status quo and California Model scenarios across each year. As recidivism declines and employment rises, total costs fall while productivity increases. The model therefore captures both reduced spending and increased economic activity.

Sources of Public Costs Associated with Recidivism

Recidivism Declines as Programming Expands

Recidivism, and all the associated social costs outlined in the previous section, is the central cost driver of incarceration. When people return to prison, costs repeat. When fewer people return, those costs decline. 

In California today, recidivism for the relevant population sits around 22 percent, measured as the share of individuals who return to prison for a new offense within a defined period after release. This definition aligns with reporting from the California Department of Corrections and Rehabilitation (CDCR), which distinguishes returns to prison from broader measures such as rearrest or reconviction.

Those broader definitions often produce significantly higher rates, with national studies showing recidivism exceeding 40 to 60 percent when measured by rearrest. By focusing only on returns to prison, this model begins with a relatively low and more conservative baseline. 

Recidivism is modeled to decline from ~22% to ~8% over ten years as education, workforce training, and reentry planning scale.

The next structural assumption in this model is that recidivism will decline gradually over the same ten-year period. With expanded education, workforce training, and reentry planning, recidivism is projected to fall steadily, tapering toward approximately 8 percent by 2035. That figure aligns with the current rate observed among returning citizen alumni of The Last Mile.

The change does not happen immediately. Instead, it phases in over time as more individuals pass through San Quentin’s rehabilitation-focused environment.

This gradual decline interacts directly with the first structural assumption: more people will be released through San Quentin each year. Given the increase in the number of releases from San Quentin, we may see the absolute number of individuals returning to prison increase in the earlier years. Still, our model projects that the rate of those returning to prison will decline over time.

Projected Recidivism Rates Over Time

Employment Rates Drive the Economic Engine of the Model

Employment is the strongest predictor of successful reentry. Individuals who secure stable work are far less likely to return to prison. They also generate wages, tax revenue, and long-term economic productivity. 

In the status quo system, employment rates for returning citizens remain low. Approximately 25 percent secure stable employment. That means three out of four people leaving prison return to their communities without consistent income, limited access to career pathways, and few opportunities to build long-term stability. 

Unemployment increases the likelihood of housing instability, financial stress, and reengagement with the conditions that often led to incarceration in the first place. When employment remains low, the system absorbs repeated costs: higher recidivism, greater reliance on public services, and lost wages that never circulate through local economies.

Employment among returning citizens is projected to increase from ~25% to ~75%, fundamentally changing post-release outcomes.

The model assumes that expanded education, job training, and employer partnerships increase employment gradually. As more individuals pass through San Quentin’s programming, employment rises toward levels already observed in structured reentry programs, an estimated 75% by 2025. This trajectory mirrors outcomes observed in structured reentry programs, including in the returned citizen population of The Last Mile. 

In the status quo system, wages for returning citizens tend to remain low, reflecting limited access to stable, career-track employment. In this model, baseline wages are assumed to average around $30,000 to $35,000 annually, consistent with the types of entry-level or unstable work many individuals secure after release.

From that starting point, wage quality is projected using two upward trajectories. A conservative scenario assumes wages rising toward roughly $55,000 annually, reflecting access to more stable, middle-income roles. An optimistic scenario assumes wages approaching $75,000, consistent with placement into higher-skilled, career-track positions. These figures represent averages across a population that includes both entry-level roles and more advanced technical careers.

The model treats wages as stable once employment begins. A person earning $35,000 in 2026 is assumed to continue earning that amount in later years. This assumption keeps the model conservative, as real wages typically increase over time in real-world work environments.

Projected Income Trajectories (Conservative vs. Optimistic)

Cumulative Productivity Creates Accelerating Savings

The most important feature of the model is cumulative productivity. Individuals who return home and remain employed generate wages every year they stay in the workforce. Each new cohort adds to that total.

In this model, someone released in 2026 contributes to the economy through 2035 if they remain employed. As employment rates rise and wages increase, cumulative productivity accelerates. By the later years of the model, productivity from earlier cohorts combines with new releases, producing a steep upward curve in total economic output.

This cumulative effect becomes the largest contributor to overall savings. The model shows productivity gains exceeding $130 million annually in later years under optimistic assumptions. The later years dominate the break-even calculation because multiple cohorts contribute simultaneously.

Projected Cumulative Productivity (Legacy System vs. New Model)

Break-Even Timeline: When the Investment Pays for Itself

The final step in the model compares cumulative savings against the initial $239 million investment in the San Quentin Rehabilitation Center. 

Rather than focusing on a single year, the model tracks how improvements in recidivism, employment, and wages translate to long-term savings for the state. Each of these changes produces measurable financial effects: Fewer people returning to prison reduces incarceration costs, more people working increases tax revenue, higher wages expand economic productivity. Together, these factors create annual savings that compound year after year.

The trajectory unfolds in phases. In the early years, costs rise slightly as more people are released through San Quentin while recidivism is still declining. This transition reflects the structural shift in how the facility is used. More individuals are being prepared for release, which temporarily increases the number of returns even as outcomes improve. During the middle years of the model, recidivism continues to fall and employment rises, stabilizing costs. By the later years, multiple cohorts are working simultaneously, and cumulative productivity accelerates. At that point, savings begin to grow rapidly.

Under conservative wage assumptions, cumulative savings approach the initial investment around 2034. Under more optimistic employment outcomes, the break-even point arrives slightly earlier, near the end of 2033. After that point, projected savings continue to expand, meaning the transformation begins generating net positive fiscal impact.

The model suggests that the transformation of San Quentin does not produce immediate savings. Instead, it gradually reshapes outcomes. As releases increase, recidivism declines, employment expands, and productivity compounds across cohorts. Over time, those shifts offset the initial investment and convert prison transformation into a long-term economic benefit.

Projected Cumulative Savings and Break-even Timeline

A Conservative Model With Expansive Implications

The model outlined here does not assume dramatic overnight transformation. It does not The model outlined here does not assume dramatic overnight transformation. It does not rely on perfect implementation, universal success, or immediate fiscal return. 

Instead, it follows a structured, phased approach in which improvements take hold over time. Recidivism declines gradually over a decade rather than dropping immediately. Employment increases steadily as more individuals move through programming and into the workforce. Wages are modeled in defined bands to reflect a range of outcomes, from entry-level roles to higher-skilled careers.

The model also accounts for a transitional period in which costs increase before they fall, reflecting the reality that more people will be released through San Quentin while the new approach takes hold. These assumptions are designed to reflect how systemic change unfolds in practice. They do not minimize the scale of potential impact, but they do ensure that the results are driven by measurable shifts in outcomes rather than idealized conditions.

Even within those constraints, the structural shift remains clear. As more individuals pass through a rehabilitation-focused pre-release environment, the number of people returning to prison begins to decline. At the same time, employment rises and wages generate measurable economic activity. 

Those outcomes build across cohorts, creating cumulative productivity that grows year after year. The model therefore captures not only the direct fiscal impact of reduced reincarceration, but also the broader economic contribution of individuals who return home prepared to work and remain in their communities.

Cumulative savings from reduced recidivism and increased employment are projected to offset the $239M investment by 2033–2034.

Investing In the Future Makes Compounding Effects

The break-even timeline emerges from that compounding effect. Under conservative wage assumptions, cumulative savings approach the original investment around 2034. Under more optimistic but still plausible employment outcomes, the break-even could occur even sooner. 

This model does not rely on aspirational claims about rehabilitation. It follows measurable inputs: releases, recidivism, employment, and wages. It accounts for both costs and benefits, including those often overlooked in discussions of incarceration. It assumes incremental progress and allows time for outcomes to mature. Within those parameters, the analysis suggests that transforming San Quentin into a rehabilitation-centered facility has the potential to shift incarceration from a recurring public expenditure into a long-term economic contributor.

If those assumptions hold, the implications extend beyond a single facility. A prison designed to prepare individuals for reentry changes the flow of people returning to communities. Fewer returns to prison reduce public costs. More individuals entering the workforce expand economic activity. Over time, those changes reinforce one another, creating a cycle defined not by repeated incarceration, but by sustained opportunity.

Our Model By The Numbers

By Robert Roche, VP of Marketing, and Sam Bufe, Sr. Manager of Research and Analytics at The Last Mile.

Want articles like this one in your inbox? Subscribe to The Last Mile Marker. This is a biweekly newsletter offering in-depth insights, critical updates, and inspiring stories on criminal justice reform and second chances.