Follow the money. The American nonprofit sector represents $2.6 trillion in annual activity, according to the Urban Institute. The Workforce Innovation and Opportunity Act alone channels $3.6 billion per year into workforce training. Federal, state, and local governments spend hundreds of billions more on housing, education, health, and social services. Philanthropic foundations add tens of billions on top. The money is not the problem. The money has never been the problem. The problem is that the architecture through which that money flows is structurally incapable of producing the outcomes everyone claims to want.
Three structural failures explain why. They aren't failures of intent. They aren't failures of effort. They're failures of design -- baked into the way funding moves, the way results are measured, and the way providers operate relative to each other. Until those three failures are addressed, more money will produce more of the same.
Structural Failure #1: Funding Cycles Create Programs, Not Systems
Grant funding operates on cycles -- typically twelve to thirty-six months. An organization identifies a need, writes a proposal, wins the grant, hires staff, delivers services, reports outcomes, and either gets renewed or doesn't. This cycle has been the backbone of American social investment for decades. It is also, by design, incapable of producing lasting infrastructure.
A twelve-month grant creates a twelve-month program. The program has a start date and an end date. Staff are hired on temporary contracts. Participants are recruited, served, and graduated. Reports are filed. When the grant ends, the program ends. The Urban Institute has documented that approximately 70% of grant-funded programs do not survive beyond their initial funding cycle. The remaining 30% often survive only by winning a new grant -- resetting the clock, rewriting the proposal, rehiring staff.
This isn't a flaw in the system. This is the system. Funders want measurable, time-bound investments with clear deliverables. Organizations compete to provide exactly that. The result is a continuous cycle of programs that start, run, report, and end -- with no mechanism for the investments to accumulate into permanent infrastructure. Every new grant is a new beginning. Nothing compounds.
Compare this to how private industry builds. A company doesn't launch a customer service department on a two-year pilot and then dismantle it. It builds the department, iterates on it, integrates it into the operating structure, and improves it over time. The infrastructure persists because it's designed to persist. The social sector builds on sand and then wonders why nothing stands.
A twelve-month grant creates a twelve-month program. When the grant ends, the program ends. Nothing accumulates. Nothing compounds. Every new grant is a new beginning.
Structural Failure #2: Reporting Measures Activity, Not Outcomes
The second structural failure is in how results are measured. The standard grant report asks: How many people were served? How many hours of training were delivered? How many certificates were issued? How many participants completed the program? These are activity metrics. They measure what the organization did. They do not measure what changed.
An organization can report that it served 500 people, delivered 10,000 hours of training, and issued 400 certificates -- and none of those numbers tell you whether anyone's life actually improved. Did the training lead to employment? Did the employment last? Did wages increase? Did housing stability improve? Did the community's economic conditions shift? Those are outcome metrics, and the funding system is not designed to collect them.
There are practical reasons for this. Outcome measurement is expensive. It requires longitudinal tracking -- following participants for twelve, twenty-four, thirty-six months after program completion. It requires data sharing between organizations that don't typically share data. It requires honest accounting of what didn't work, which is politically uncomfortable for both the funder and the funded organization. So the system defaults to activity metrics because they're easy to collect, easy to report, and easy to celebrate.
The consequence is that funders and policymakers have almost no reliable data on what actually works. They have mountains of data on what was attempted. They can tell you exactly how many people sat in a classroom. They cannot tell you, with any consistency, whether those people's lives are different because of it. Decisions about where to invest next year's billions are made on the basis of activity reports from last year's programs -- a feedback loop that optimizes for volume, not impact.
Pew Research Center data shows that only 22% of Americans trust the federal government to do the right thing. That trust deficit isn't mysterious. People see the money being spent. They see the programs being announced. They don't see the results in their neighborhoods. The gap between spending and outcomes is visible to everyone except, apparently, the people writing the checks.
Structural Failure #3: Providers Compete Instead of Coordinate
The third structural failure is the most damaging and the least discussed. In any given American city, the social service landscape is fragmented across dozens of independent organizations, each operating its own programs, maintaining its own data systems, and competing for the same limited funding. Brookings Institution research has documented that a mid-size city can have 30 or more workforce and social service providers operating in the same geography with no coordination mechanism.
This isn't a coordination problem. It's a competition problem. The funding model creates it. When thirty organizations are competing for ten grants, the incentive structure rewards differentiation, not collaboration. Each organization needs to demonstrate that its approach is unique, its model is proprietary, its outcomes are superior. Sharing data with another provider means potentially sharing credit -- or worse, revealing that another organization is producing better results.
The human cost of this fragmentation is staggering. A single family in need of services might interact with a workforce training provider, a housing assistance organization, a food bank, a childcare support program, and a healthcare navigator -- five separate organizations, five separate intake processes, five separate case files, five separate reporting systems. None of these organizations know what the others are doing. The family is the only entity that holds the complete picture of their own service engagement, and they have no structured way to communicate it.
A city might have 35 workforce providers. None sharing data. All competing for the same grants. This isn't corruption. It's architecture.
I've seen this in Detroit. I've seen it in Cleveland, Atlanta, and Baltimore. Good organizations doing genuine work, staffed by people who care deeply about their communities -- and all of them operating in isolation because the funding model punishes coordination and rewards competition. A workforce provider that shares its curriculum with a neighboring provider risks losing its competitive edge in the next grant cycle. A housing organization that shares its outcomes data with a partner might reveal that its placement rates are lower than the partner's. The rational organizational behavior, within the current funding architecture, is to protect your data, protect your model, and protect your funding relationships. The result is thirty organizations all doing versions of the same work, none of them learning from each other.
The Architecture Problem
These three failures -- short-cycle funding, activity-based reporting, and provider competition -- aren't independent problems. They're interlocking features of a single architecture. The funding cycles create the programs. The activity metrics justify the programs. The competition fragments the programs. Together, they produce an ecosystem that consumes enormous resources and produces minimal systemic change.
This is not a corruption story. Nobody is stealing the money. The people administering the grants, designing the programs, and delivering the services are overwhelmingly competent and well-intentioned. The architecture is the problem. The incentive structures, the measurement frameworks, the competitive dynamics -- these are design choices that produce predictable outcomes. And the outcome they produce is churn: a continuous cycle of new programs replacing old programs, new reports replacing old reports, new announcements replacing old announcements -- while the underlying conditions remain largely unchanged.
The $3.6 billion in annual WIOA spending is a clear example. That money flows through state workforce agencies to local boards to training providers. At every handoff, administrative costs accumulate, reporting requirements expand, and the connection between the dollar and the outcome it's supposed to produce grows more attenuated. By the time that federal dollar reaches a person who needs job training, it has passed through three or four organizational layers, each taking a percentage for overhead, each adding reporting requirements, each optimizing for its own institutional survival.
What Would Have to Change
The fix is not more money. The fix is different architecture. Funding cycles would need to extend to five or ten years, with built-in iteration rather than annual recompetition. Measurement would need to shift from activity metrics to outcome metrics tracked over multi-year horizons. And the competitive model would need to be replaced with coordinated systems -- shared data infrastructure, unified intake processes, collaborative outcome tracking -- where organizations are rewarded for contributing to system-level results rather than defending organizational turf.
None of this is technically difficult. The data systems exist. The coordination models exist. The outcome measurement frameworks exist. What doesn't exist is the institutional willingness to adopt them, because adoption would require every player in the current architecture -- funders, intermediaries, providers, government agencies -- to give up some measure of control, visibility, or competitive advantage.
The money is there. It has always been there. The coordination isn't. And until the architecture changes, the next $2.6 trillion will produce the same results as the last.
