August 2026
The Opportunity Capital Investment Fund
Scaling Programs that Create Economic Mobility Using Private Capital
The Problem
Many Americans are looking for better, higher-paying jobs but lack clear, affordable paths to reach them—a longstanding challenge that is likely to grow as AI reshapes the labor market.
Some workforce training and higher education programs do create pathways to upward mobility by helping people build skills that employers need. But even the most effective programs remain too small to meet national demand.
The central barrier is financing: people seeking training or education often cannot afford to pay up front and cannot easily borrow to do so. As a result, high-performing programs remain dependent on philanthropy and public funding, limiting their ability to scale even when they could increase earnings, business productivity, and long-term economic growth.
Executive Summary
Opportunity Capital is an investment fund designed to help scale workforce training and other economic-mobility programs that increase people’s earnings.
Investors provide up-front capital to expand effective programs and are repaid by the government using a share of the additional income tax revenue generated when participants earn more.
The near-term goal is to launch a pilot in 2027 and, if successful, develop broader legislation to enable scaling the financial model and the creation of a new data-driven market for opportunity.
The Opportunity Capital Solution
Opportunity Capital uses a new evaluation method enabled by big data to create a way for private investors to fund programs that raise workers’ earnings and earn returns:
- Investors provide funds to Opportunity Capital, which uses the funds to expand programs that have a proven record of increasing workers’ earnings.
- Workers enroll in these programs (without having to pay for them) and earn more money upon completion.
- The Census Bureau measures the earnings gains and resulting additional federal income tax revenue using a new Digital Twins machine learning algorithm developed by Opportunity Insights.
- The government pays a share of that additional tax revenue back to investors, creating an expected annual net return of about 10% with little volatility, at no additional cost to workers.
Digital Twins: A New Method of Measuring Program Impacts
The key advance that makes this model feasible is a new Digital Twins methodology to measure program impacts developed by Raj Chetty and collaborators at Opportunity Insights (Chetty et al. 2026).1 In simple terms, the method estimates what each participant would likely have earned without the program by comparing them with similar people of the same age, gender, location, and prior earnings and employment history using a machine learning model. The model is trained on anonymized tax and Census data covering 170 million American workers. Program impact is then measured as the difference between participants’ actual earnings and those predicted benchmark earnings.
Importantly, Chetty et al. (2026) show that this algorithm produces estimates very close to gold-standard randomized trials. Unlike randomized trials, however, it can be run quickly and at very low cost using existing data across many programs, making annual measurement of earnings gains and tax revenue feasible at scale.
What Returns Will the Fund Generate?
Our analyses show that a typical highly effective workforce training program costs about $10,000 per participant and increases earnings after enrollment by roughly $9,000 per year on a sustained basis. This earnings increase generates $1,700 in additional federal income tax revenue each year.
10%
average annual return with near-zero beta.
If investors received all of the incremental tax revenue for twenty years, a $10,000 up-front investment would produce around $34,000 in total cash flows (measured in 2026 dollars), equivalent to a 16.7% gross nominal internal rate of return.2 In practice, we expect governments to retain a portion of the added tax revenue and investors to bear fund management expenses, yielding net rates of return around 10%.
Because payouts depend on earnings gains relative to what participants would otherwise have earned, expected returns are relatively stable over the business cycle and have little correlation with stock market movements (near zero market beta).
While investors earn significant returns, the social returns to their investment are even larger: for every $1 earned by private investors, workers gain $10 over 20 years.
Piloting Opportunity Capital Investments
Implementing the proposed financial model at scale will require legislation that enables federal or state governments to pay a share of measured tax revenue gains to private investors. As a step toward that goal, we are working to pilot the model using two federal programs: the Department of Education’s FIPSE Postsecondary Student Success Grant Program and the Department of the Treasury’s Social Impact Partnerships to Pay for Results Act (SIPPRA). In these pilots, investors will be repaid an amount equivalent to the additional federal income tax revenue generated by participants’ higher earnings, using funds that have already been appropriated for workforce innovation and outcomes payments. We have applied for $24 million of funding from FIPSE and expect to apply for approximately $20 million from SIPPRA.
The federal funds in these programs must be paid out within five to seven years by statute. To expand the scale and scope of the pilots so they can capture earnings gains over a longer time horizon, we are supplementing the federal grants with matching philanthropic capital. Blue Meridian Partners and the Surgo Foundation are committing $15 million for a 1:1 match of federal dollars. We are currently seeking $29 million of additional philanthropic funding to fully match the anticipated $44 million of federal funding we hope will be available for the pilot, yielding a total pool of $88 million for outcomes-based payments in 2032 (equivalent to $70 million in today’s dollars, discounted at the 10-year Treasury bond rate).
We plan to use the $88 million pool of funds to repay investors five years after their initial investment (based on projected earnings gains over a twenty-year horizon at that point). The $88 million pool of funds in 2032 will be able to support capital investments of approximately $50 million in 2027 given an expected return of 10%. We have preliminary commitments of $15 million from investors thus far and are currently seeking an additional $35M of investment capital to launch the fund in 2027.
Payments from philanthropic matching funds will be made in proportion to the realized federal income tax revenue gains. If programs do not deliver earnings gains, philanthropies pay nothing. In this sense, the model effectively creates a mechanism for impact-based philanthropic funding, with private investors bearing up-front risk and philanthropies only paying when outcomes are achieved.
Philanthropies interested in expanding this model beyond matching federal dollars could do so by pledging additional resources. Philanthropies could also participate on the investment side of the model via program- or mission-related investments from their own portfolios.
Advantages Over Previous Pay for Success Initiatives
The financial model proposed here addresses two key challenges that limited earlier pay-for-success efforts such as social impact bonds and income-based loan repayment. First, the Digital Twins method eliminates the need for expensive randomized trials and thereby allows ongoing measurement of program impact. Second, because investor repayment is tied to realized tax revenue gains, the model creates a continuing source of capital rather than a one-time demonstration without requiring the government to commit any funding up front.
Critically, the Digital Twins method ties returns to earnings gains from program enrollment, unlike traditional income-based repayment programs that are based on earnings levels alone. This encourages investors to back programs that truly add value, not simply those that enroll participants who were likely to earn high incomes anyway. In this respect, the model is similar to Tax Increment Financing (TIF), which gives developers a share of the added tax revenue generated by real estate and infrastructure investments. Our proposal is effectively TIF for people, using income tax revenues combined with precise measures of what participants would have earned otherwise.
Long-Term Vision
A successful pilot of the Opportunity Capital fund will pave the way for legislation to create a much larger capital market for opportunity in the United States. Such legislation could occur at the federal or state level, using income tax revenues to support variants of the model described above. Preliminary conversations at both the federal and state level – ranging from Massachusetts to Texas to Utah – have generated substantial bipartisan interest.
Creating a market for opportunity would have transformative impacts. In 2025, the United States spent roughly $800 billion on higher education and workforce training, often with limited evidence on economic impact. Redirecting even a small share of that spending toward more effective programs using the evidence-based financial model proposed here could create benefits for many stakeholders: better jobs and upward mobility for workers, more funding for effective education and training providers, more productive talent for employers, a new source of returns for investors, lower transfer spending for taxpayers, and ultimately greater economic growth for the United States.
Identifying Effective Programs
The fund’s success hinges on identifying cost-effective programs that increase workers’ earnings.
To date, we have measured returns for programs run by Per Scholas, Year Up United, Project QUEST, Western Governors University, the Massachusetts Higher Education system, Dallas College, and Texas State Technical College. We are seeking additional data partnerships to expand the scope of training and higher education programs considered for investment, following protocols established by Opportunity Insights to share data in a manner that preserves privacy of participants and adheres to all applicable regulations.
Team
Opportunity Capital LLC is led by Raj Chetty, William A. Ackman Professor of Economics at Harvard University. Chetty is among the most widely cited economists of this generation and has pioneered the study of economic mobility using big data as the Director of Opportunity Insights. Opportunity Capital is supported by several partners. Opportunity Insights, a non-profit research center based at Harvard University, is leading work on methodology and identification of program impact. Social Finance, an investment advisor with extensive experience in public-private partnerships for social impact, is leading work on contracting between the private and public partners. Seward and Kissel LLP has been retained to prepare fund documents and term sheets. Opportunity Capital is additionally supported by an advisory board with expertise in financial management, workforce training, and pay for success programs.
Interested in learning more? Please contact us at info@opportunity-capital.com
- Source: Chetty, R., Fogel, J., Katz, L., Noray, K., Porter, S. & Reisinger, J. (2026). Digital Twins: Evaluating Workforce Training Programs in the Age of Big Data. Working Paper.
- The 16.7% internal rate of return reflects the annualized compound return on the initial investment if the annual cash flows—beginning two years after investment and growing at 2.5% per year with inflation—are reinvested in a portfolio with that same rate of return.
Methodological Appendix
Digital Twins
A New Method of Measuring Program Impacts
Measuring Program Impact
There are thousands of workforce training and similar programs in the U.S. that aim to increase earnings for participants, but we have limited evidence on which programs are effective. The most reliable evidence to date comes from randomized trials, which provide credible measures of program impact but are costly to implement. Chetty et al. (2026)1 have developed a new Digital Twins method that produces estimates of program impact that are comparable to those from randomized trials, does not require a resource-intensive evaluation, and can be scaled at virtually no additional cost.
The Digital Twins Method
The key step in measuring a program’s impact on earnings is identifying what each participant would likely have earned without the program—their counterfactual earnings.
The Digital Twins method uses a machine learning model to estimate this counterfactual by comparing program participants with similar people of the same age, gender, location, and prior earnings and employment history. The model is trained on non-participants using anonymized tax and Census data covering 170 million American workers. Program impact is then measured as the difference between participants’ actual earnings and their predicted counterfactual earnings. The Digital Twins method improves on earlier efforts to measure program impact without a randomized trial by using large-scale administrative data rather than smaller survey samples. These larger and richer data can fit more flexible models that incorporate a wider range of information and more accurately predict counterfactual earnings.
Experimental Validation
Importantly, the Digital Twins method produces program impact estimates very close to those from randomized trials, which use a randomly assigned control group to measure counterfactual earnings. Figure 1 compares estimates of program impact on earnings from randomized trials and the Digital Twins method across 37 randomized trials of workforce programs. The points cluster around the 45-degree line, and the two sets of estimates have a correlation of 0.93, indicating close agreement between the Digital Twins and randomized-trial estimates of program impact.
The Digital Twins method yields estimates comparable to those from randomized trials—the gold standard in program evaluation—without requiring changes to program implementation. Its ability to measure program impact using only a list of participants is especially valuable because earnings impacts of workforce programs can vary widely across sites and enrollment cohorts, even within a single training provider.
Interested in evaluating your program using the Digital Twins method?
We are actively seeking additional data partnerships. Opportunity Insights ensures that the data sharing process protects participants’ privacy and complies with standard regulations. Reach out to data@opportunityinsights.org for more information.
- Chetty, R., Fogel, J., Katz, L., Noray, K., Porter, S. & Reisinger, J. (2026). Digital Twins: Evaluating Workforce Training Programs in the Age of Big Data. Working paper.
