[
California produces more economic output than all but four countries. It also has water mains that predate the New Deal, highways that routinely score D+ from civil engineers, and a housing shortage so severe that tent encampments line the streets of cities with trillion-dollar economies. The state that builds the future can’t maintain the present.
This isn’t a “California is bad at everything” thing. This defining condition of American infrastructure in the 2020s across most states, red or blue. Broadband speeds increase; bridges deteriorate. Private capital builds data centers in months; public agencies spend years on environmental review for a bus lane. The federal backlog of deferred maintenance, estimated by the ASCE’s 2025 Report Card at a $3.7 trillion investment gap over the next decade, continues to compound. Crumbling roads, failing water systems, outdated transit, overwhelmed stormwater drainage; few public works perform as they should.
It is a bit of a generalization. But the pattern is real, and its costs fall hardest on people least equipped to absorb them: the commuter on the bus that never comes, the family drinking water from corroded pipes, the neighborhood that floods because the drainage was designed for a climate that no longer exists.
Something has to change. On that, nearly everyone agrees. What to do about it is another matter.
One name keeps surfacing: the U.S. Army Corps of Engineers (but there are more CoEs like the Navy that get’s overlooked!) The Corps builds levees, dredges harbors, constructs military facilities, and responds to disasters with a speed that makes other agencies look paralyzed. It is, by wide consensus, the most operationally effective infrastructure institution in American government.
Which raises a question that virtually everyone concerned with infrastructure eventually asks.
We understand the impulse, and it reflects something real. The Corps is remarkably effective. When Congress authorizes a project, the Corps mobilizes with a discipline and velocity that state agencies struggle to match. If that effectiveness stems from identifiable practices (management techniques, procurement strategies, organizational design) then replication seems straightforward. Study the model. Transfer the lessons. Build faster.
Tempting, but wrong. Or rather, importantly incomplete.
The question rests on a misunderstanding of why the Corps is effective, one that leads reformers to focus on the wrong variables and overlook the structural forces that actually drive the gap between federal and state capacity. Understanding that gap, its sources, its size, etc, is the central task for anyone serious about improving American infrastructure. The gap is narrower than defeatists assume. It is also wider than optimists acknowledge. And it has specific, identifiable causes that suggest specific, actionable responses.
But to see those responses clearly, we first need to understand what makes the Corps work. The full picture is more interesting, and more instructive, than the usual accounts suggest.
The standard explanation starts in the right place: federal sovereignty. The Corps wields the Supremacy Clause not as an abstract legal principle but as an operational tool.
The great Carlos Mucha will point out that whenever Congress authorizes anything from housing to a levee, local zoning ordinances don’t apply. The Corps doesn’t negotiate with county planning commissions or submit to local design review. It builds what Congress authorized, where Congress authorized it, according to federal standards. Period.
A state DOT trying to widen a highway through a suburb faces years of hearings, environmental litigation, and competing local ordinances. The Corps building a flood control channel through that same suburb operates in a different procedural universe.
The Corps is not exempt from NEPA. It conducts environmental assessments and writes impact statements. But it occupies a structurally exceptional position under the statute, one that explains why states struggle to replicate its speed.
Most federal agencies that touch infrastructure (the Federal Highway Administration, the Federal Transit Administration) are reviewing other people’s projects. A state DOT proposes a highway; FHWA reviews the environmental documents. The proposing entity and the reviewing entity are separate institutions with separate incentives, and the back-and-forth adds months or years to every project. The Corps is different: for its own civil works projects, it is simultaneously the proposer and the lead federal agency for NEPA review. Planning, environmental review, engineering design, and construction management sit under one roof. No other major federal infrastructure agency combines these functions so completely.
The Corps is not entirely self-policing. It must consult with the Fish and Wildlife Service, coordinate with the EPA, and comply with the National Historic Preservation Act and other statutes (33 CFR 230.9). But the institutional integration means the Corps controls the pace of compliance in a way external applicants never can.
States encounter NEPA from the outside, as applicants needing federal permits or funding recipients who must satisfy a federal agency’s review. A state water agency that needs a Section 404 Clean Water Act permit must submit to the Corps’ own review, an irony not lost on officials who find themselves waiting on the very institution they are trying to emulate. The state controls neither the timeline nor the process, and the reviewing agency has every institutional incentive to be thorough, which in practice often means slow.
Recent legal shifts have reinforced this advantage. The Supreme Court confirmed in 2025 that NEPA is procedural, not substantive: agencies must analyze environmental impacts but need not reach any particular outcome. The Fiscal Responsibility Act of 2023 imposed enforceable time and page limits on reviews. Then in July 2025, the Corps overhauled its own NEPA procedures, expanding categorical exclusions, allowing them to be stacked to cover entire projects, and narrowing the scope of required analysis for most permit actions.
The upshot: local opponents can comment on a Corps project. They cannot block it through NEPA alone. And the Corps, unlike a state agency standing in line for federal review, controls the tempo of its own compliance.
There’s more. The Corps enjoys the federal “navigation servitude,” a dominant property right over navigable waters rooted in the Commerce Clause. When the Corps dredges a channel, it often doesn’t have to compensate property owners for altered water flow. States navigate takings clauses and compensation requirements at every turn, adding cost and time to every project that touches private property.
And then there’s liability, or the absence of it. Section 702c of the Flood Control Act of 1928 provides that “no liability of any kind shall attach to or rest upon the United States for any damage from or by floods or flood waters at any place.” If a federal levee fails, the government is generally immune. States get sued. When Oroville Dam’s spillway failed in 2017, forcing the evacuation of 188,000 people, California faced over $1.1 billion in repair costs. A federal agency would have walked away, pushing the costs to FEMA or some other emergency federal government program. That asymmetry shapes every engineering decision, every risk calculation, every timeline.
Sovereignty, navigational dominance, liability immunity. No state can match this through better management alone.
But here we need to complicate matters.
Does sovereignty alone explain the Corps’ performance?
No. That’s only a part of the answer.
Sovereignty alone doesn’t build anything. Plenty of federal agencies enjoy legal privilege and still perform poorly. HUD has sovereign immunity; nobody holds it up as a model of operational excellence. The VA operates under federal supremacy; its construction projects are notorious for cost overruns and delays. The Aurora, Colorado medical center, budgeted at $328 million, has exceeded $2 billion in total expenditures, so badly mismanaged that Congress stripped the VA of authority over large construction projects and gave it to the Corps. The federal government’s legal advantages are necessary for the Corps’ effectiveness. They are plainly not sufficient.
The Corps works because it combines constitutional privilege with management competence. We need both halves of the story.
Start with the hybrid military-civilian structure. Military officers command districts and divisions on three-year rotational tours. This rotation, imposed by the Army’s personnel system, prevents the territorial fiefdoms that calcify in agencies with permanent leadership. A district commander who knows she’s moving in 36 months has every incentive to deliver results and none to build a local empire. Meanwhile, a civilian workforce of roughly 36,000 provides institutional memory, technical expertise, and the long-duration relationships with Congress and stakeholders that military rotators cannot maintain. Military officers provide dynamism; civilians provide depth.
Then there’s concentrated expertise. Rather than expecting every district to master every discipline, the Corps runs Centers of Expertise. Portland houses the Hydroelectric Design Center; Walla Walla standardizes cost estimation nationwide. This architecture means fifty dam safety engineers working together catch errors that isolated specialists going to miss. When New Orleans is overwhelmed by hurricane recovery, it can offload design work to other centers without loss of quality. And when individual experts retire, the center retains their methods and accumulated judgment.
The training infrastructure reinforces this. PROSPECT, the Proponent-Sponsored Engineer Corps Training program, delivers over 200 courses annually to some 7,000 employees, training tailored to Corps missions and taught by Corps personnel. A Corps engineer follows a structured, career-long curriculum.
Finally, enforced standardization. The Army Facilities Standardization Program maintains standard designs for 70 facility types across nine Centers of Standardization. When Fort Hood needs new barracks, the Corps adapts an existing model rather than designing from scratch. A base commander cannot demand unique architectural flourishes. The standard is the standard.
We’ve built the model up as strong as we can make it. Now for the hard part.
Because three structural barriers stand in the way.
States cannot legislate supremacy over their own municipalities the way the federal government asserts supremacy over states. The legal relationship between states and localities runs through two competing doctrines: Dillon’s Rule (localities possess only those powers expressly granted by the state) and Home Rule (localities have inherent authority over local affairs, often constitutionally protected). Most major American cities operate under some version of Home Rule. When a state agency wants to build through a Home Rule city, courts frequently apply balancing tests: is this a matter of statewide concern or local concern? If judges deem it a local land-use question, the state agency must comply with local ordinances, facing the same delays as a private developer. The federal government skips this balancing entirely. The Supremacy Clause establishes hierarchy.
Then there’s scale. The Centers of Expertise model requires a project portfolio large enough to keep specialists continuously engaged. The Corps builds dams, levees, navigation channels, and military facilities across all fifty states. That continental scope justifies a Dam Safety Modification Center staffed with dozens of engineers holding doctorates. A single state, even California, doesn’t build enough dams to sustain that concentration of talent. The Corps’ national portfolio smooths the peaks and valleys. States, operating within their own borders, cannot.
Finally, culture. The Engineer Regiment is a military institution. Officers rotate because the Army rotates officers. They accept postings to undesirable locations because the Uniform Code of Military Justice governs their careers. They execute with urgency because military culture treats delay as failure. You cannot transplant this into civilian state agencies. Civil service rules, union contracts, and democratic accountability create fundamentally different incentive structures. A state DOT director serves at the pleasure of the governor, navigates legislative micromanagement, and manages a workforce with employment protections designed (for good or for ill) to prevent exactly the kind of command authority that military organizations take for granted.
That’s the disillusionment. Here’s what states can actually do.
The standard debate tacitly assumes the Corps model is uniformly desirable. That assumption deserves scrutiny.
Consider liability immunity. Section 702c shields the federal government from flood damage claims. We’ve described this as an advantage, and operationally it is: engineers freed from liability exposure can make faster decisions without defensive over-engineering. But immunity from consequences is not the same thing as accountability for outcomes. When a federal levee fails, affected communities have limited legal recourse. Is that just? The question is genuinely difficult. Liability exposure slows states down, but it also forces a responsiveness to affected populations that the federal model lacks. States might reasonably decide that some version of liability, capped, structured, predictable, is preferable to full immunity, even at the cost of speed.
Or consider the navigation servitude. The Corps’ ability to alter waterways without compensating property owners is legally powerful. It is also, in communities where waterfront property represents generational wealth, a source of legitimate grievance. States that lack this tool face higher costs. They also face their constituents more honestly.
The gap between federal and state capacity is partly a function of features states shouldn’t want, not just features they can’t have. Once we see this clearly, the task shifts. We’re not trying to replicate the Corps. We’re trying to identify which dimensions of the gap can be narrowed through strategies consistent with state-level governance, and which dimensions represent genuine, defensible tradeoffs.
Three state-level experiments illuminate what’s possible and where the ceiling is.
Florida’s Water Management Districts represent the most ambitious state-level attempt at Corps-like capacity. Five districts, created by the Water Resources Act of 1972, with a distinguishing feature that matters more than any other: constitutional taxing authority. The Florida Constitution authorizes WMDs to levy property taxes directly, up to capped millage rates. The districts don’t compete for annual legislative appropriations. They have their own checkbook.
This enables planning horizons that appropriations-dependent agencies cannot match. The Comprehensive Everglades Restoration Plan, a $26 billion effort spanning decades, requires sustained investment across multiple gubernatorial administrations. A normal state agency, dependent on the annual budget cycle, could never commit credibly to that timeline. The WMDs can, and do. South Florida WMD owns and operates over 2,175 miles of canals and 2,130 miles of levees, exercises eminent domain for water storage, and partners with the Corps not as a supplicant but as an equal. The July 2025 Everglades agreement saw Florida take over construction of the EAA Reservoir inflow and outflow pump stations, a nine-pump facility capable of moving 3 billion gallons of water per day, because the state could move faster than its federal partner, accelerating the project timeline from 2034 to 2029.
When a state agency can credibly say “we’ll do it ourselves if you’re too slow,” the entire dynamic shifts. Fiscal autonomy doesn’t grant sovereignty. But it grants leverage, and leverage, in intergovernmental relations, is worth a great deal.
The limitation is real. Florida’s arrangement is constitutionally entrenched and emerged from specific geography: the state is, to a first approximation, a flat wet platform where water management is existential. Replicating this in states where infrastructure needs are more diffuse would face enormous political resistance. Voters and legislators are rarely enthusiastic about creating new taxing authorities.
Fiscal autonomy can substitute for sovereignty in specific domains. Sadly, it can’t generalize easily.
If Florida tests whether money can substitute for sovereignty, California tests whether expertise can. The California Department of Water Resources manages the State Water Project, the largest state-built water conveyance system in the world. Thousands of engineers and scientists. Enormous institutional knowledge. By raw technical capacity, DWR looks like a mini-Corps.
But DWR operates in a regulatory environment that systematically negates the advantages of scale.
CEQA, the California Environmental Quality Act, requires detailed environmental review and allows extensive citizen litigation over Environmental Impact Reports. The procedural requirements are genuine and serve legitimate purposes. They have also become delay tools for opponents who dislike projects for reasons unrelated to environmental protection. Neighborhood groups have used CEQA to block homeless shelters, bike lanes, and affordable housing; one study found that 80% of CEQA lawsuits target infill development rather than genuinely harmful projects. CEQA litigation delays housing projects by an average of 2.5 years. And unlike the Corps, which houses planning, engineering, construction, and permitting under one roof, DWR negotiates with sister agencies at every stage. The State Water Resources Control Board regulates DWR’s own water rights. Interagency permitting adds years to timelines that are already long.
After Oroville, California faced $1.1 billion in repair costs and over $1.7 billion in damage claims filed by farmers and businesses, exposure that a federal agency would have avoided entirely because of FEMA (who chose a small partial reimbursement to California) and liability caps. While useful for the private sector, liability warps public sector engineering judgment. The question becomes not “what happens when this fails” but “who gets sued when this fails,” and the answers push in different directions, because Oroville still failed and California lost money that it could used on “what happens when this fails”. The first produces robust design. The second produces defensive design, gold-plated documents, and risk tolerance approaching zero.
Consider the counterexample. In the 1970s, the mayor of Fudai, a small village on Japan’s northeast coast, insisted on building a 15-meter floodgate across the town’s inlet. Residents and officials called it excessive. He had survived a tsunami as a child and refused to budge based on what happened, actual risk, in the past. When the 2011 Tōhoku tsunami hit, Fudai was one of the few coastal communities in the region that suffered no deaths and minimal damage. The floodgate held. That was engineering driven by the question “what happens when this fails,” pursued by someone with the authority and the will to act on the answer. California has the engineers. It doesn’t have the legal room to let them think that way.
If Florida and California each test one dimension of the gap, Texas tests a different strategy: rather than building a single Corps-like institution, Texas borrows federal authority where possible and deploys financial engineering where it can’t.
The signature achievement is NEPA Assignment. Under federal law, the Federal Highway Administration can assign environmental review responsibilities to state DOTs. TxDOT assumed this authority in 2014 and now writes its own Environmental Impact Statements and signs its own Records of Decision for highway projects. For those projects, TxDOT effectively becomes the federal agency. In July 2025, Texas renewed this arrangement for ten years, double the previous term, a reflection of sustained competence and persistent advocacy.
This cuts years off project timelines by eliminating the federal-state back-and-forth that plagues highway construction elsewhere. Texas takes on legal liability (a genuine cost) but gains speed (a genuine and, in Texas’s judgment, sufficient benefit). The authority doesn’t extend to water projects or navigation, but within that domain, the value is real.
On water, the Texas Water Development Board operates as an infrastructure bank, using the state’s credit rating to issue bonds and lend cheap capital to local water suppliers. And the Texas Legislature regularly preempts local ordinances that conflict with state infrastructure priorities, asserting state supremacy within its constitutional authority rather than deferring to local veto points.
Targeted borrowing and strategic preemption can narrow specific dimensions of the gap, even if no single strategy closes it entirely.
None achieved anything resembling full replication. All achieved meaningful improvement in specific domains. The right approach is not a single Corps-like institution but a portfolio of strategies, each targeting a specific dimension of the gap.
No single strategy closes the gap, but targeted interventions can narrow each dimension of it. The recommendations below address the three structural barriers, sovereignty, scale, and culture, plus mechanisms that bypass the gap entirely.
The most direct attack on the sovereignty gap is aggressive zoning preemption. Most states have some statutory preemption for state-owned facilities, but the exemptions tend to be narrow and contested. States can expand legislatively, declaring infrastructure projects “matters of statewide significance” and explicitly preempting local zoning, building codes, and design review for designated project categories. This will not deliver federal-style supremacy, courts will still apply balancing tests, but it reduces veto points meaningfully. The political trade is important to frame correctly: the state takes on liability and cost; in exchange, it controls siting and design.
Beyond preemption, states should pursue NEPA Assignment wherever federal law permits. Texas proved the model works for highways. Other states, particularly those with large federal-aid highway programs, should seek delegation aggressively, staffing dedicated offices to manage the assignment relationship and maintain compliance. Texas is currently one of just seven states with active agreements; its ten-year extension in 2025 required demonstrated competence and sustained advocacy over years. States that invest in that capacity now will be positioned to expand delegation as Congress extends the model to additional project types.
These are preconditions. Sovereignty-narrowing measures should come first, because everything that follows works better once veto points are reduced.
The Centers of Expertise model is the Corps’ most transferable organizational innovation, but it requires adaptation. A single state can’t replicate the Corps’ continental portfolio. It can, however, designate specific offices as State Centers of Expertise for disciplines where knowledge concentration matters most: hydraulic modeling, alternative delivery contracting, environmental compliance, asset management. Staff these with career specialists. Require consultation for all relevant projects statewide. Make them mandatory repositories for institutional knowledge.
The staffing question is where most proposals go vague. It shouldn’t. State Centers of Expertise need a permanent core of engineers and technical staff whose institutional memory doesn’t walk out when a contract expires. They own the models, the data, and the standards. But sustaining that core requires a support structure that a single state agency typically lacks.
Formalized university partnerships, under state direction, can provide it. Universities offer research capacity, a graduate student pipeline, and analytical depth in specialized disciplines. Alaska DOT’s partnership with the University of Alaska Fairbanks on 2D hydraulic modeling guidance shows the right pattern: state agency leads, university supports. For multi-state compacts, the MAGNET model among Alabama, Georgia, and Mississippi, three governors, three state universities, and private partners creating a shared R&D hub, shows how university consortia can anchor compacts that would otherwise lack institutional permanence.
Models, data, and decisions stay with the state (or, in the case of constitutionally autonomous entities like Florida’s Water Management Districts, with the body that holds mission authority). Universities get access, just not custody. A professor’s incentives run toward publication and grants. A state center’s incentives must run toward operational reliability and project delivery. When ownership is ambiguous, the academic incentive wins by default.
The key is institutionalization. Ad hoc task forces dissolve when the crisis that created them passes. University MOUs lapse when the grant cycle turns. Centers persist when their existence is codified in statute, their funding is baselined, and their consultative role is mandatory.
For disciplines where even a large state can’t sustain sufficient concentration, multi-state compacts offer a path to Corps-level scale. Gulf Coast states could jointly pre-qualify levee contractors and share specialized inspection capacity. Western states managing shared river basins could negotiate joint procurement for common project types. These arrangements are administratively complex. They are also, for certain specializations, the only realistic option.
States should also consolidate purchasing power. The Corps’ procurement advantage is fundamentally about volume. States can aggregate demand through cooperative purchasing agreements, pre-qualify contractor pools for common project types, and allow localities to issue task orders against master contracts rather than running full independent procurements.
Military culture can’t be transplanted into civilian agencies. But the culture gap is really two distinct problems, and conflating them leads to imprecise solutions.
The first is federal-facing: state personnel don’t understand how the Corps works, which makes them poor partners. The fix is straightforward. Some Corps PROSPECT courses already accept non-federal participants. States should systematically send engineers through this pipeline as a structured professional development requirement. Engineers who attend absorb Corps methods, build relationships with federal counterparts, and bring institutional knowledge home. It also builds the personal relationships that make intergovernmental partnerships work.
More broadly, states should invest in partnership capacity: staff who understand WRDA authorization processes and cost-sharing formulas, technical personnel capable of evaluating Corps designs rather than accepting them on faith, fiscal systems that can meet matching requirements without legislative delay. Florida’s WMDs partner with the Corps as equals because they made these investments deliberately. The Corps prefers such partners and moves faster when working with them.
The second is internal: state agencies struggle to attract and retain technical talent, and the specialists they do employ often work in isolation rather than in the knowledge-dense environments that make Corps centers effective. This requires a different tool.
States should enact talent exchange statutes, equivalents of the federal Intergovernmental Personnel Act, authorizing bidirectional rotations between state agencies and universities. Currently only two states match the federal IPA’s scope in clearly permitting such exchanges. A state engineer who spends a year in a university hydraulics research program returns with sharper technical skills and current knowledge of emerging methods. A professor who rotates into a state dam safety program brings analytical depth and returns to campus with operational understanding that reshapes what she teaches. Over time, this thickens the bench that State Centers of Expertise draw from without requiring permanent headcount increases.
Workarounds are important. Empowering special districts is one. Texas’s Municipal Utility Districts, special-purpose governments that issue bonds, levy taxes, and build infrastructure, solve problems that defeat delivery through conventional channels. A developer creating a new subdivision establishes a MUD, which finances and constructs infrastructure before homes go up. The model has real drawbacks: high tax burdens on early residents, opaque governance, fragmented regional coordination. But MUDs deliver dedicated revenue outside legislative appropriations battles and infrastructure built ahead of demand rather than years behind it.
Infrastructure Benefit Districts offer another mechanism. When a state builds a highway interchange, surrounding property values rise, often substantially. Define a district around a project, establish baseline assessed values, dedicate a share of future appreciation to infrastructure. As the area develops, enabled by the very infrastructure the district finances, the funding stream grows. This doesn’t require sovereignty, scale, or military culture. It requires legislative authorization and political will.
States should also facilitate consolidation of utility districts and similar organizations responsible for public infrastructure and utilities, not by mandating it (which generates fierce resistance) but by creating strong incentives. Metropolitan areas often contain dozens of jurisdictions with separate infrastructure responsibilities. The Corps thinks in watersheds; localities think in city limits. States can offer bonus matching funds for multi-jurisdictional projects, streamline joint powers authority creation, and condition permits or grants on demonstrated regional coordination.
We return to the question we started with, because we can now answer it properly.
The Corps is fast because it combines a history of operations know how with constitutional privilege, sustained work on a continental scale and underwritten by a military culture that civilian institutions cannot reproduce. States can’t replicate that combination wholesale. The barriers are structural features of American federalism, not problems of ambition or leadership.
But the gap has identifiable dimensions, and each admits of specific responses. Fiscal autonomy can substitute for sovereignty. Borrowed federal authority can eliminate years of procedural delay. Multi-state compacts can approximate scale advantages. Training pipelines can close knowledge gaps. And mechanisms like special districts, benefit districts, and liability caps can achieve similar outcomes through entirely different institutional channels.
Think of a San Francisco street, of self-driving taxis navigating potholes, the cutting-edge office building overlooking tent encampments. The gap between private dynamism and public stagnation is real, and it is not closing on its own. But the path forward does not require states to become something they constitutionally cannot be. It requires them to pursue, with discipline and specificity, the considerable improvements that lie between wholesale replication and passive acceptance of decline.