Contracts, Forecasts, and the Silence of Warning Signals: What Really Collapsed Arizona's CAP Irrigation Districts
Operational Analysis | Water Infrastructure Governance
Contracts, Forecasts, and the Silence of Warning Signals: What Really Collapsed Arizona's CAP Irrigation Districts
E-Score Water · April 2026 · 9 min read
The collapse of Arizona's CAP irrigation districts in the early 1990s was not an inevitable consequence of geography, drought, or bad luck. It was the predictable outcome of decisions made decades earlier — decisions to commit to infrastructure costs that economic analysis had already shown were unaffordable, to suppress warning signals that challenged a politically important project, and to lock district finances into contractual structures with no escape provision when reality diverged from projection.
In January 1994, the Central Arizona Irrigation and Drainage District became the first federally supported irrigation district in U.S. history to file for bankruptcy. A second filing followed in August. Neither event was a surprise to anyone who had read the right documents. The surprise was that the warnings in those documents had been systematically set aside for nearly three decades.
The Project That Political Will Built
The Central Arizona Project was authorized by the Colorado River Basin Project Act of 1968 — the culmination of decades of lobbying by Arizona's congressional delegation. The engineering rationale was sound: by mid-century, Arizona was pumping groundwater at 500 times the rate of natural recharge, aquifers were measurably collapsing, and growth demanded a new supply. Construction began in 1973.
The 336-mile aqueduct was designed to deliver 1.5 million acre-feet of Colorado River water annually into central and southern Arizona, with agriculture projected to absorb the largest share while municipal demand grew into its entitlements over time. Irrigation districts in Pinal, Maricopa, and Pima counties were enrolled as primary subcontractors. They borrowed money to build distribution systems. They signed take-or-pay contracts. They were locked in.
What they were not given was an honest accounting of whether the underlying economics would ever work.
Two Studies, Two Conclusions — One Buried
Two distinct analytical frameworks were applied to the question of whether Arizona farmers would find CAP water economically viable. The Bureau of Reclamation used an ability-to-pay framework: could farmers theoretically generate enough revenue from crops to cover CAP costs? Structured optimistically, the answer was yes.
Young and Martin (1967) and Kelso, Martin, and Mack (1973) used profit maximization analysis: given the actual cost of CAP water relative to available alternatives — including cheaper groundwater already in the ground — would a rational farmer choose to buy it? The answer was no. CAP water would not be economically viable for Arizona farmers under realistic market conditions.
The profit maximization studies were not obscure academic exercises. They circulated within the planning community. Their conclusions were known. The decision to proceed on the basis of the ability-to-pay framework was a deliberate choice in service of a project that had powerful political backing and that Arizona desperately wanted. The method was selected because it produced the required answer — not because it was more rigorous.
Feasibility Studies Built on Four Flawed Assumptions
University of Arizona economist Paul Wilson systematically documented the structural errors in the late-1970s and early-1980s distribution system feasibility studies — the analyses used to support federal loan approvals and private district bond issuances. He identified four categories of failure:
- 100% acreage utilization assumed. The studies treated all CAP-eligible acres as farmed every year. This was not true before CAP and did not become true after delivery began.
- Crop mix radically overstated. Studies projected a shift to high-value vegetables that had no basis in historical planting patterns. The lettuce projection alone exceeded documented reality by a factor of ten.
- Groundwater substitution ignored. A farmer with access to cheaper pumped water will use it first and treat surface water as a supplement. This obvious economic behavior — the rational default for every grower in the service area — was absent from the analysis.
- No uncertainty modeling. Commodity prices, yields, and marketing channels were entered as point estimates with no sensitivity analysis, no probability ranges, and no downside scenario. A single bad assumption anywhere in the chain invalidated the entire projection.
These were not minor methodological quibbles. They were the load-bearing assumptions on which hundreds of millions of dollars in district debt was issued. When those assumptions failed to materialize — and they failed almost immediately — the districts were left holding contracts with no financial headroom and no renegotiation path.
CAP Agricultural District: Decision Timeline & Failure Sequence
The Take-or-Pay Trap
The mechanism of collapse was not simply that CAP water cost more than groundwater — that was always true. The mechanism was the take-or-pay contract structure, which required agricultural districts to pay a fixed rate on all water allocated to them annually, whether they used it or not.
When water began flowing in 1985, municipal and industrial demand across the Phoenix and Tucson metro areas was well below original projections. Because irrigation districts held the residual allocation — whatever remained after M&I and Indian priority uses were satisfied — the districts found themselves contractually responsible for over one million acre-feet per year and approximately 75% of CAP's $30 million annual fixed operations, maintenance, and replacement costs.
By 1992, the Bureau of Reclamation's own regional economist had documented in a memo that enforcing the take-or-pay provision would result in most districts defaulting on their federal distribution system loans and being forced into bankruptcy. The agency knew. The structural flaw had been internally identified. The districts were nonetheless placed into formal repayment status.
Three Failure Patterns Utility Managers Should Recognize
The CAP irrigation district collapse is most often framed as a story about Western water politics or Colorado River overallocation. That framing is not wrong, but for operational water utility managers it is the least actionable part of the story. The operationally instructive version is about three institutional failure patterns that are neither unique to Arizona nor to the 1970s.
Pattern 1 — Feasibility Analysis That Serves the Project Rather Than Tests It
When a feasibility study exists to support a decision already made, it will systematically exclude the assumptions most likely to produce negative conclusions. The ability-to-pay framework was not neutral analysis — it was a method selected because it produced the required answer. Any utility entering capital commitments or bond issuance should ask: what assumptions, if wrong, would make this economically nonviable? Has anyone stress-tested those independently?
Pattern 2 — Warning Signals That Are Institutionally Inconvenient
The Young and Martin study was not ignored because it was poor analysis. It was sidelined because it was inconvenient analysis. Organizations that suppress contrary findings do not eliminate the underlying risk — they accumulate it silently until it materializes as a crisis. The districts that declared bankruptcy in 1994 cited a 27-year-old study in their court papers. The institution had known the whole time.
Pattern 3 — Contractual Lock-In Without Demand-Side Flexibility
Take-or-pay provisions are common in large water infrastructure financing. The CAP case illustrates what happens when they are layered onto revenue projections built from optimistic demand forecasts with no downside scenario. Any utility entering long-term supply or treatment contracts should model performance under a scenario where actual usage runs 30–50% below projection. If the contract has no adjustment mechanism and the utility has no reserve capacity, essentially all demand risk has been transferred onto the utility with no offsetting protection.
The Resolution Confirmed the Diagnosis
The eventual resolution — codified in the 2004 Arizona Water Settlements Act — required irrigation districts to relinquish their long-term CAP entitlements entirely in exchange for debt relief and access to a smaller, affordable fixed volume of lower-priority water. The relinquished entitlements were reallocated to tribal and municipal users who could make the economics work.
In other words: the resolution required undoing the demand allocation framework in place since the 1960s, writing off substantial federal and private debt, and rebuilding the priority structure from scratch. A decade of litigation, bankruptcy proceedings, and federal negotiations to arrive back at something approximating the economic reality that the profit-maximization studies had predicted in 1967.
Boring Analysis Beats Heroic Infrastructure
The CAP canal is a genuine engineering achievement. Its 336-mile aqueduct delivers water to over 1.8 million people in the Phoenix metro area and remains the backbone of Arizona's renewable surface water supply. None of that diminishes what happened to the irrigation districts that financed the system's agricultural buildout.
The infrastructure worked. The analytical and contractual frameworks surrounding it failed — and they failed in ways that were documented in advance, ignored for institutional reasons, and correctable at multiple points before the bankruptcies occurred.
Water utilities contemplating large capital investments, long-term supply contracts, or major demand projections would benefit from applying a single discipline: find the most unfavorable credible economic analysis of the project and ask whether the organization could survive if that analysis proved correct. If the honest answer is no, the contract structure, financing approach, or demand assumptions need to change before the commitment is made.
Arizona's irrigation districts in 1994 could not answer yes to that question. They could not answer it because nobody in a position of authority had asked it honestly since 1967.
Sources: This analysis draws on University of Arizona Water Resources Research Center documentation; Bureau of Reclamation project records; CAP Library Agriculture History Report (October 2016); W. Michael Hanemann, "The Central Arizona Project," UC Berkeley Working Paper No. 937; Paul Wilson, University of Arizona economic assessments; and the Justia record of Central Arizona Irrigation and Drainage Dist. v. Lujan, 764 F. Supp. 582 (D. Ariz. 1991). Published by E-Score Water — escorewater.org.
