Levers of power: uneven adjudication and conflict resolution over long-standing property rights to water

Despite rapid advancements in infrastructure, economic growth, technology, and agricultural science, concerns over water have been front-of-mind in the western United States since early American expansion in the 1800s. The natural aridity of the western states required surface water diversions for mining, development, and agriculture, the last of which involved large-scale irrigation to maintain crops that required significantly more water than was naturally available over wide regions.

To sustain a growing population, diversions were necessary to bring water across miles of desert throughout the west. Legal institutions developed early on for claiming and maintaining recognized rights to divert and use water. These rights developed alongside many of the geographic and climactic realities of the region — water had to be diverted from its source and transported to its point of use due to the scarcity of streams and rivers. Legal institutions developed to protect diverters’ rights as hefty investments were required to put water to use (Leonard and Libecap, 2019).

These institutions developed both formally and informally, balancing realities and needs in local areas within the overarching purview of the federal government, which relied on a growing local western populace for both development, and also local management of industry and rights for land and resources. Before the 1900s, much of the investment required for constructing diversion and irrigation came from private prospectors or early irrigation companies. But the infrastructure required substantial amounts of money invested into harsh environments and for uncertain outcomes, and yet the federal government underwent a concerted effort to induce settlement from the east to the western frontier. Thus at the turn of the twentieth century, as western expansion began to boom (and private irrigators went bankrupt), the US federal government became actively involved in providing public funds for private development via the Bureau of Reclamation ^[1], as it pushed to populate the west.

If the system of apportioning water were as simple as matching institutions to patterns of use and scarcity, the story of western water use would have been a simple one. But two important factors influenced how western law and institutions shaped western development vis-à-vis water use, and vice versa: the existence of American Indian tribes throughout the western frontier, and their contracts with the United States over land and resources (including water) dating back to the early-to-middle 1800s; and the evolution of federalism in terms of state versus federal rights as they related to land and resources in and around the public domain. Many early conflicts over water in the west dealt with balancing federal and state interests, or in protecting or expropriating rights and resources from American Indian tribes by way of muddled and evolving federal and state power. In this paper, I examine the underlying shifts in power between these entities and how they are expressed in the context of water use and rights.

For example, many western American Indian tribal nations had been corralled and kept by force on federal reservations dating back to the California gold rush. To help keep tribes contained within these reservations, which were often in arid and inhospitable regions for conducting agriculture or sustaining development, the U.S. government entered into contracts (treaties) with tribal nations delineating reservations and terms of agreement. The federal government often promised resources, infrastructure, supplies, and personnel, in addition to money from property sales, in exchange for a tribe ceding land and territory. Many of these treaties were entered into within territories that later entered the union as states if they were not already. As external population pressures increased, however, these tribal reservations soon became natural targets for resource expropriation, particularly due to the fact that tribal rights and resources were controlled by the federal government, who, through representative government, was amenable to local pressure via Congressional action in appropriating federal resources for investments in infrastructure, or using federal statute to support western settlement and development.

In essence, where state power intruded upon that of the federal government and tribal nations, it was done so by federal allowance. This balance of power between states, tribes, and the federal government has adjusted over time, sometimes swinging more towards states’ rights and at others protective of tribes. It is typically the federal government that has legally adjusted how this power balances, but on-the-ground action can (and has historically) served as levers of power in rebalancing this trifecta relationship over time.

In fact, many conflicts between these three entities, and especially between tribal nations and local settlers (often acting under state law and jurisdiction), involved water. Because tribal land and rights were held in trust by the federal government, many Native nations had little recourse in the late 1800s and early 1900s for protecting water resources that were contractually promised, or established via caselaw or statute, from being usurped by encroaching settlers outside or even within their reservation borders. Layer on top of that competing jurisdictions over land and potentially water as states entered the union, and very quickly water rights were being appropriated faster than the law could keep up.

The turn of the twentieth century saw these issues rise to the U.S. Supreme Court (USSC) to weigh in on how to protect water rights embedded in treaties between the U.S. government and tribal nations from pressures from new users appropriating water under state laws. In what has become the seminal tribal water rights case, U.S. v Winters^[2], the USSC ruled essentially on how to balance state versus federal power — i.e., whose rights had preeminence.

While presently Winters is often referred to as defining tribal rights to water, the court held that the U.S. government had reserved rights to water for the tribes of the Fort Belknap Indian Reservation, and that power of the United States to reserve such water “could not be” [$…$] “denied ^[3].” With this ruling, the justices prioritized the U.S. government’s treaty with tribes over the states’ appropriation of water to settlers. This established what would become known as the Winters doctrine, and is a clear elaboration of state appropriation laws’ subordination to federal power. The court explicitly ruled on preeminence of federal power in the context of American Indian federal reservations, recognizing the federal government’s inherent reservation of water in establishing tribal reservations. This was an implied right, and was not subject to later-instituted state laws governing process and use more locally.

While this was a clear legal delineation, in practice, these conflicts were far from over. The rationale for this was twofold. Firstly, the federal government, in this ruling, declined to set a mechanism for quantifying these water rights. Yet at the same time, they deferred the allocation of water rights to states. Without explicit boundaries around what had to be reserved for tribes, states allocated water elsewhere. Secondly, using water in the west is a capital-intensive operation, strongly tied to laws that support first-user advantage for typical state appropriation schemes. If the law claimed that tribes held vast quantities of reserved water in theory, on-the-ground water users (often supported by state laws and even federal funding) could establish and strengthen de facto rights to use water and then fight in courts or lobby lawmakers to refine institutions to recognize these practices de jure. These tactics often grew from the ground up, staking a line in the sand in practice and then pushing the levers of power through water appropriation and economic development, to then influence institutional change.

It is important to consider that tribal water rights were expressly tied to reservation land, and that early evolution of the Winters doctrine also occurred alongside active diminishment of federal reservation land, through legal means. Between 1880 and 1915, over 100 million acres of reservation land was transferred out of tribal ownership or status to largely support broader American western settlement, largely by Congressional statute or federal jurisprudence (see Carlson, 1981; Taylor, 2018). In terms of water, and in this context, tribes had few practical avenues to construct, implement and maintain irrigation infrastructure without it being usurped for other federal or state purposes, thanks in large part to the trust-trustee relationship between tribes and the United States. As western encroachment continued, tribal land that was irrigated was often leased and then ceded to western settlers through federal and state mechanisms, leaving the Winters right poorly enforced.

Thus what were early conflicts between competing users for water resources became entrenched conflicts that persisted — and still persist — for decades if not over 100 years in some cases. Due to tribes’ trust relationship with the U.S. government, until relatively recently they had little recourse to enforce or protect these important rights. In 1975, the federal government passed the Indian Self Determination Act ^[4], which enabled tribes greater autonomy in contracting for services and in practicing sovereignty in governmental action. This brought a sea-change of legislation and policy shifts that enabled tribes to pursue settling tribal water conflicts to have Winters rights quantified. In recent decades tribes have begun in earnest — and with success — to engage in processes to enforce, quantify and practically put Winters rights to use.

Practically speaking, for tribes to engage in these processes, however, they must go through a series of costly and arduous tasks to either litigate or negotiate for the enforcement and quantification of their rights to water, to transact over these rights, and to practically use these rights how they see fit. Congress must ratify all components of litigated or negotiated outcomes, and often provide funding where needed for implementation. The process itself is inherently political with major legal changes influencing how and through which courts tribes can take action.

As of 2020, out of more than 200 tribes that ostensibly have rights to reserved water, there have been 81 settlement negotiations and/or litigation processes on behalf of 80 tribes; with 25 ongoing. As a result, tribes now hold senior use rights to about 20% of water in the entire Colorado River Basin. Many more tribes have negotiations in process and many others have not yet initiated their claims for rights. These cases represent repeated examples of conflict resolution over water, and over time.

In this paper, I seek to understand the process of starting these procedures and resolving them through the lens of the bigger-picture shifts in balance of power between the federal and state governments, and tribal nations. Thanks to recent economic research, we understand that environmental characteristics such as soil quality, rainfall and drought, in addition to economic conditions such as agricultural activity impact which tribes initiate, and then are able to settle, reserved water rights cases. Sanchez et al. (2020) found that increasing water scarcity, and increasing competition for resources (illustrated by increasing off-reservation population pressures) catalyze disputes, but the more parties to the negotiation delay resolution. In addition, several papers have looked at the impact of resolving or clarifying property rights on some outcome (Browne and Ji (2023) and Deol and Colby (2018) relating to American Indian tribes and agricultural revenue). Sanchez et al. (2023) also show that settlements increase cultivated acreage.

Yet there are also other important structural and dynamic factors which influence the outcome and trajectory of these cases from a political perspective. I present an examination of “Winters” cases in the early years following the 1908 ruling and on through the mid-to late twentieth century, to present the important institutional shifts in the balance of power between states and the federal government. This is one of the contributions of this paper to the literature, and is an often overlooked period in modern analysis of Winters rights. I also research the development of de facto use of water rights by competing settlers, outlining the interplay between on-the-ground use and enforcement and the changing legal context. The through-line through all of this is the protection or allowance of state jurisdiction into tribal and federal rights.

I then add to this historical examination a first-pass empirical analysis to tease out whether even in broad brush strokes and measurements we see patterns that illustrate and help explain the political relationships. I investigate the probability of initiating Winters rights, and then examine observable and time-varying factors that influence not only if, but when, a tribe may initiate proceedings. All of these frameworks include institutional shifts as correlating factors. From this perspective, the use of water is inherently a lever of power enabling rights or disabling others’, depending on what complementary infrastructure or institutional factors are at play. This is an important part of the story, and is still one that influences the several outstanding and potential water rights settlements today.

This paper is organized as follows: The next section outlines early use of establishing de facto water use rights in order to supersede legal rights that were unenforced. Sections “State versus Federal Rights” and “Assimilationist Water Policies” then discuss early cases in developing the Winters doctrine and key shifts in legal precedent that has allowed the balance of power to tilt toward the state in terms of tribal water rights. Empirical analyses are presented in the sections “Understanding Enforcement and Quantification” and “Empirical Analysis”, and the final section concludes.

Over the past 100 years, the federal government has both asserted, but often did not actually protect, tribal rights to water in the American west. Deferring the management and allocation of natural resources largely to state and local governments has allowed those entities the ability to shape, restrict, or expand tribes’ rights in practice through on-the-ground management. I present an institutional analysis of the evolution of this dynamic, and how the Winters doctrine evolved as such, and then present empirical findings on the factors that influence which tribes are able to negotiate for and settle rights that have long been recognized in theory but not in practice.

The most basic limitation to supporting and enforcing tribes’ rights to water in the twentieth century was the appropriation of this water to other users. Even what we now describe as the “seminal” case protecting water rights was technically a case about federal power, and practically a case fought for non-Indian water users leasing tribal land. In that seminal case, U.S. v Winters (1908), no American Indian was actually involved directly in the filing, negotiating, procuring or planning for the litigation, and the tribes’ role in the process was seen as indirect by the Office of Indian Affairs (Shurts, 2000).

In earlier attempts to secure water rights, one superintendent of the reservation had filed for tribal water rights in his own name (Luke Hays in 1898), and another opened up much of the irrigable reservation land to grazing leases for non-Indian stock farmers, including himself (William R. Logan), and to develop a sugar beet industry using tribal water. Tribes on the Fort Belknap reservation were routinely dissuaded from using water to support a stock industry (the more lucrative venture) in favor of agriculture, which in turn just supported non-Indian stock farmers. Revenue from leased reservation land was collected on behalf of the tribe by the federal government, and then spent on Indian wage-labor in support of the non-Indian industries, whose proprietors leased the land.

Almost instantly after the 1908 Winters decree, nearby off-reservation users, supported by the U.S. Department of Interior and the State of Montana, began undoing Winters through the harsh “realities of capital flows (Shurts, 2000)” — i.e., by largely siphoning investment in water development projects to off-reservation users and not to the tribe. That same year, the U.S. Bureau of Reclamation took 2,587 acres of irrigable land from the Fort Belknap Indian Reservation to construct the Dodson Dam and its canal ^[5]. In the years that followed, local off-reservation irrigators successfully lobbied for the construction of several additional water diversion and conveyance projects from the US government, including the Nelson Reservoir, the Glacier Park and St. Mary Canals, and the Sherburne, Vandalia and Fresno dams (Wolfe, 1992). Meanwhile, the Fort Belknap Indian Irrigation Project (the oldest federal Indian irrigation project in the country) ^[6], began construction in 1889 but was never completed (FBIC, 2021), and the Fort Belknap Indian Community still has no congressionally-ratified quantification standard upon which to base their legal rights to this day.

Even in cases where the Office of Indian Affairs wanted to secure water for the actual tribes, the method for doing so was unclear given conflicting and evolving laws between state and federal governments. In 1880, an Office of Indian Affairs agent with the San Carlos Agency implored, “[i]If there is any law in regard to this it should be enforced, so that the Indians can be protected in their water rights, a matter of vital importance [$…$] ^[7]” Before, the Commissioners’ reports outlined several tactics to try and legally assert rights, from leasing land to white cattlemen and farmers (in order to put water to use), to allotting land in order to assert individual property rights and use water through individual farms (which was often difficult, particularly since relatively little money was available for irrigation infrastructure to tribes, especially before 1900). Many plans documented in the historic reports also indicate even if a scheme to retain water or build water infrastructure was devised, it often was not put into action due to either lack of funds, or because the water had already been used surrounding the reservation (United States Office of Indian Affairs,1880-1920).

Where the federal government did expend money for water infrastructure, it was often for non-Indian purposes, or to keep American Indians contained on reservations. As early as 1867, the federal government began funding the irrigation of Indian lands, understanding that in order to contain American Indians on newly-created reservations — many of which were in desert areas, they would need to invest in procuring water. The initial round of funding constituted $76,000, and then in 1890, after 9,000 American Indians left the Navajo reservation in search of water, the federal government expended more (Pisani, 1986).

Other irrigation projects implemented by the federal government, constructed at a time when irrigation investment otherwise came from private sources, included the building of infrastructure using funds from cessions of tribal lands. The Crow Nation, for example, entered into an agreement in late 1890 with the U.S. to cede 1.8 million acres of reservation land in exchange for $946,000 (an average valuation of $0.525 cents per acre). The contract allowed for substantial portions of those funds to be spent on irrigation infrastructure and related machinery and labor ^[8]. By 1896, over $250,000 was expended on such investments, with 80% of that money spent on labor (65% of which was for tribal laborers), canals, ditches, pipe and other public works ^[9]. In the annual reports for the Office of Indian Affairs, the local agents routinely wrote of the excellent farming land on the Crow reservation, and expended substantial funds from Crow Nation reserves on building irrigation works ^[10]. Over a million acres of Crow land was ceded again in the early 1900’s ^[11].

Yet in the ensuing years, after substantial amounts of money was expended on the project (sourced from Crow funds managed by the U.S. federal government), many of the larger irrigation projects petered out in investment, and significant acres of irrigated land were in actuality used by non-Indian users (Voggesser, 2001). On the Blackfeet reservation, also in Montana, early Indian agents described land that was ill-suited for agriculture, but great for mining, ceding this land to non-Indian settlers and concentrating the tribe on would-be agricultural plots. For an illustration of quickly-changing boundaries, juxtaposed against competing interests for water, Figure 1 shows the Blackfeet and Crow reservations between 1890 and 1915. In both cases, reservations are diminished, and non-Indian federal irrigation projects are established often just next to or near current reservations — setting up potential for future conflict, or present conflict over built resources in the early 1900s.

As in the Milk River and Crow cases, many tribal laborers were often paid from proceeds of tribal land sales — turning what was meant to be tribal profits for aggregate land sales into small wages paid to tribes on hourly bases (Pisani, 1986). In some cases, the Department of Interior made tribal members labor to construct irrigation ditches that would never be filled ^[12], turning wealth into wages, and arguing that labor was important for character building, all the while creating wealth for land speculators and western non-Indian settlers.

While the Office of Indian Affairs did continue to fund reclamation projects, many of them were taken back for the Bureau of Reclamation projects, essentially transferring higher-value irrigated land (that tribes had — sometimes without consent — paid to irrigate out of their “surplus” land sales) directly to settlers. Figure 2 shows a broader picture of public lands reserved under the 1902 Reclamation Act (United States Department of Interior, 1905). By 1904, 22 projects were under construction or approved, and an additional 48 projects were under investigation (United States Department of Interior, 1905). Many of these lands initially claimed and reserved by the Reclamation service had recently before that been part of large Federal Indian reservations (1880 reservations depicted by red-shaded areas). In later appeals cases, courts noted that from the inception of the Bureau of Reclamation, the federal government, as a policy, “enshrined” the diversion of Indian reserved water for non-Indian use (Shurts (2000)inAnderson (2006))

By 1910, over half a million dollars had been expended or appropriated on tribal reclamation projects through federal government channels, creating over 375,000 acres under ditch (Pisani, 1986; again, much of this funded by ceded land sales paid to the federal government on behalf of respective tribes). This is in addition to the fact that proceeds from public land sales, including American Indian reservations, funded the reclamation service, and such sales “largely increased” after the passage of the Reclamation Act (Office of Indian Affairs Annual Report, 1906, p. 42; see Figure 3). Pisani (1986) points out that this represented, at the time, nearly half the irrigable land within BoR projects. Twenty-four Reclamation Service projects were approved by 1906, without the land to construct these projects (Pisani, 1986). They turned to reservations, relying on federal law that enabled the Reclamation Service access to any land within a reservation, as long as they provided a small individual farm in return (Pisani, 1986).

Some estimates place that by 1912, eight million dollars had been invested by the federal government for irrigation projects, investment that was “almost wholly [dependent] on the legal rights to use of that water ^[13].” Shurts (2000) reports that a second memorandum (unsigned and unpublished) was also written by the Office of Indian Affairs in around 1913 or 1914, describing litigation it was preparing for to protect tribal rights from encroachment of non-Indian users and/or state jurisdictional conflict. In this memorandum alone, at least 12 cases/conflicts are listed across the west, mostly regarding encroaching users drawing on tribal water (Shurts, 2000, p. 196).

The Office of Indian Affairs also contended that not only were there influxes of encroachers physically claiming water and making legal claims to it as well, but the Reclamation Service was much better prepared to represent those interests against those protecting tribal rights. Shurts (2000) describes that in that second Office of Indian Affairs memo, the bureau described how non-Indian claimants had some of the best, well-prepared council, while the government defending tribal rights was “frequently if not always caught unprepared.” Non-Indian claimants were also often private irrigation companies, armed and ready for legal fights. The Reclamation service also was much more staffed – consisting of “eight or nine Examiners, one for the office of each Supervising Engineer in charge of a district, with a general office division of seven under the Chief Law Officer with a Supervising Engineer in direct charge ^[14],” which was a stark contrast to the significantly fewer personnel of the Indian Office (Shurts, 2000).

Thus, by the time the Ninth Circuit, and then the U.S. Supreme Court ruled on the Winters case, there was already substantial investment in development of irrigated land both on and off reservations. In fact, the defendants of the Winters case had populated the northern border of the Fort Belknap reservation on the Milk River through public domain land entry acts, and invested in diversion infrastructure to divert water to put under beneficial use under Montana state law. They argued they had no notification of tribal rights to be aware of and that when Montana became a state, it repealed reservation of Milk River water that had been established under treaty rights (Nelson and Booke, 1977). The court ruled a state could not curb powers enjoined by the federal government to “reserve and exempt” waters from appropriation under state law (Nelson and Booke, 1977).

The above section outlined key on-the-ground practices that inherently set up conflicts over water rights, but what about the legal side? Winters was in its essence “[a] question of jurisdiction […] presented by the United States ^[15].” Much of the legal jockeying around water rights and enforcement in the twentieth century was around jurisdictional issues between states and the federal government. Especially during the American Colonial and Constitutional periods, treaties with Indigenous nations caused the “primary struggles” between states and the national government (Ablavsky, 2015), and some saw reserved federal rights as interfering with efficient operation of prior appropriation within each state.

Recognition of reserved federal water rights inherently restricts the exercise of state sovereignty, and in some cases, preempts provisions in state constitutions or law (Peterson and Tong, 1983). The rights of states to adjudicate and manage water rights within their borders relative to the power of the federal government has ebbed and flowed over the past century. There also have existed sometimes conflicting arrangements for balance of power.

For example, 11 states (Montana being one of them) entered into the union with clauses “disclaiming” jurisdiction over Indian tribes and Indian land that were required of them by the federal government (Sommer, 1983; Wilkins, 1998) ^[16]. These clauses were designed to give assurances to tribes and the federal government that states would never, without tribal consent or federal authority, interfere with internal affairs of tribal nations.

In practice however, states often pushed the boundaries of this jurisdictional limit via litigation, but also by in-practice application and enforcement of state resource management policies, particularly as they related to water.

Many in recent years describe the trajectory of Winters enforcement and the development of the Winters doctrine as jumping from the USSC ruling in 1908 to the method proposed in 1963 for quantifying reserved rights ^[17]. But there were important moves in the evolution of the doctrine in the intervening years, particularly in ruling on the boundaries of Winters rights, and how state-defined rights could interfere (or not). Because encroachment was a pervasive reality in and around tribal reservations (and an effective method for usurping previous ownership or rights decrees), very early “Winters” cases often dealt with:

• Attempts to clarify upper boundaries on reserved rights quantities — based on previous use or potential future needs?

• Whether a reserved right was retained for non-Indian leaseholders or purchasers of allotted trust land.

Table 1 presents some of the key cases that illustrate the shifting balance of power over water post Winters. Many of these cases were about how to support non-Indian water users within the letter of the law, protecting an interpretation of Winters.

In Byers v Wa-Wa-Ne (1917) ^[18], Oregon courts narrowly interpreted Winters rights to be limited to only what was necessary to fulfill the Umatilla tribe’s treaty with the United States government. Despite the recent reserved right doctrine, Oregon allowed encroachers to divert water away from the reservation. Initially the state of Oregon (and Congress) was careful to sidestep permanent encroachment on the tribe’s reserved rights, but after enough years had passed, and the nearby economy became dependent on the water, Congress, by legal statute, allowed off-reservation users a permanent right to the water in question, at the expense of the tribe ^[19]. The Oregon Supreme Court relied on both the preeminence of Congressional statute to extinguish treaty provisions, and also the minimal need concept that would come up in later decades — i.e., that the tribe was not using their reserved water right for irrigation, so it did not “need” it (this is despite the fact that the part of the reservation that remained post cessions needed investments in improvements in order to successfully irrigate — improvements the federal government did not wish to invest in).

In the 1920s, two cases, Skeem v. United States^[20] and United States v. Hibner^[21] affirmed that water rights for tribal allotted land was not relinquished when land was leased or purchased by non-Indian users. Importantly, in Skeem, the court ruled that the amount of water tribes had reserved rights for did not have to be based on what had been currently or previous used, but based on what could be cultivated. This is a clear precursor the “Practically Irrigated Acreage” distinction in later decades. U.S. v. Hibner also addressed the issue of whether the American Indian individual allottee needed to be residing on the land for the water right to be retained. Again, the courts found that no, “failure of the Indians to use their water will not cause either an abandonment or a forfeiture of their rights thereto ^[22].” Purchasers or lessors could receive a grant of water via tribal rights, but their use would be governed by state laws. Note both of these cases which on the surface are more amenable to higher upper bounds of reserved rights quantities were asserting rulings in situations where non-Indian water users leasing or purchasing tribal land would benefit and were often the actual water users in question.

This doctrine continued to evolve over time. In United States v. Powers^[23], the U.S. Supreme Court again affirmed that reserved water rights are maintained for individual allotted parcels even after conveyance or death of or by the original allottee. In this instance, the court refused to injunct water users on former allotment land as requested by off-reservation users. United States v McIntire^[24], also affirmed in the 9th Circuit court at the same time that prior appropriation does not apply to Indian lands.

At this point, jurisdictionally, the law was clear: state rules did not apply within federal Indian land or to usurp federal reserved rights. Yet time and time again, state water users, making claims under state law, via state legislators and state representatives chipped away at boundaries of federal protection. It would only be so long until practically the federal government adjusted jurisdiction to legally allow the state to embody more control over the practical apportionment of water within their boundaries.

Two important changes brought this shift starting in the 1950s, representing a steady increase in the number of areas states could take over jurisdiction regarding previously federal matters (i.e., “assimilationist” policy in a jurisdictional sense), and initially without tribal consent. The first was the McCarran Amendment ^[25], which granted jurisdiction to state courts to hear federal water law cases. This meant that the United States could be joined to a suit for the adjudication of water rights. The impact of this amendment was significant — it essentially restricted tribes in their latitude to pursue judgements over water rights in federal courts, directing them to have to engage with state courts. In particular, to join the United States as a party in state court, it had to be through a general stream adjudication. Shortly following, in 1953, Congress passed Public Law 280, transferring criminal and civil jurisdiction over Indian lands from federal to state governments. Five states, Alaska, California, Minnesota (limited), Nebraska, Oregon (limited), and Wisconsin received mandatory transfers, and others (Nevada, South Dakota, Washington, Florida, Idaho, Montana, Arizona, Iowa and Utah) opted in without tribal consent ^[26]. Thus, it was the federal government’s expression that granted the states’ increasing authority over tribal matters in state courts.

Tribal sovereignty to a large extent, in this context, is a prevention of state law intrusion into Indian land; the federal government, through statute and caselaw, opted throughout time to dimmish or uphold protection of that intrusion as it related to water rights. As population pressures grew on western resources, the federal government facilitated deeper intrusions of state law. In the aftermath of the McCarran Amendment and PL-280 (which, in tandem, drastically reduced protections for tribal or federal jurisdiction relating to water), there was an open question as to the applicability of state law regarding tribal water rights, especially given that many of the applicable states also had “disclaimer” clauses in their enabling acts or state constitutions.

In 1976, the US Supreme Court interpreted the McCarran Amendment as authorizing the adjudication of tribal water rights in state courts, in a court case also referred to as “Akin ^[27]”. This set up a conflict between lower courts — one that established the McCarran Amendment could supersede “disclaimer” clauses, and one that claimed it could not. In Jicarilla Apache Tribe v. United States^[28], the 10th circuit court held that disclaimer states could adjudicate tribal water rights (“assimilationist” policy). Three years later, in Northern Cheyenne Tribe v. Adsit^[29], the ninth Circuit court ruled that state jurisdiction disclaimers prohibited states from assuming jurisdiction over adjudicating tribal water rights. In this guise, the McCarran Amendment cannot be read to amend a state constitution regarding jurisdiction over tribal water rights ^[30]. Notably, the 9th Circuit, in its opinion, presented concern that “Akin” could set up a situation where tribes would effectively be prohibited from ever fully adjudicating their rights in federal court, since the state could join the United States as a relevant party. The judges recognized that, as demonstrated in “Adsit,” the state had incentives to prevent and stall long enough to enact comprehensive statute that would limit the protections for tribal rights to water, to circumvent federal jurisdiction (Peterson and Tong, 1983).

These inherent conflicts were resolved a year later by the U.S. Supreme Court. In Arizona et al. v. San Carlos Apache Tribe^[31], the USSC agreed with lower courts that in fact federal water rights should be determined in state courts, regardless of disclaimer status. This reversed the 9th Circuit decision in “Adsit”, and sidestepped the concerns over practicality and enforcement. This decision held even if the adjudication only included an Indian tribe. Effectively, this closed the door on limitations on state courts enacted in disclaimer clauses, and opened the door to intrusions granted by the McCarran Amendment (NARF, 1983).

As the Native American Rights Fund (NARF) elaborated when these restrictions became clear, “[s]uch proceedings are extremely costly and take literally decades for completion. They also may be politically unpopular in a state, not only because of the expense, but because all water users in the state will also be required to prove up their rights. It is unlikely then that states or other parties will rush into state court to begin general stream adjudications (NARF, 1983).”

Other interpretations at the time pointed out the inadequacy in relying on state courts to protect tribal or treaty rights. In a letter written by attorney Anthony Rogers to Senator Edward Kennedy in 1976, Rogers stated that: “[t]he likelihood is remote that state judges will apply Indian law and federal treaty and statutory interpretation to Indian rights questions as favorably to Indians as will federal judges.” He blamed this squarely on politics:

The adherence to state courts also meant states could interpret previous doctrine as they saw fit. In many cases, for example, a lack of quantification standard allowed for a flexible or generous interpretation of how much water tribes had rights to, with an eye toward fulfilling an overall homeland purpose that could extend as future needs or development changed. Within the state of Arizona, however, the state Supreme Court has fostered how “Winters Rights” are applied, creating a doctrine where any quantification must be “practical” and not “pie-in-the-sky.” This inherently reframes discussion over rights about what is “minimal” to fulfill a pressing need — a subsistence-minded approach inherently more so than a growth-minded one. Thus, it is still quite complicated again to implement an intended reserved right so many years after determination of legal status.

Underlying, and often driving, these conflicts over federal reserved water for tribes, are issues in scarcity (how much competition is there for water), capital infrastructure, and intrusion or protection of tribal rights defined through the federal government. In the hundred years since the Winters ruling, relatively few tribes who ostensibly could claim rights under the reserved rights doctrine have been able to successfully start negotiations with de facto water users, agree to a reallocation of water, and have that agreement ratified by Congress and the required infrastructure, which in many cases may have been owed as a treaty stipulation for creating “homelands”, funded and built in order to put water rights to practical use.

As outlined earlier, resolution proceedings to enforce and quantify tribal water rights are costly, long and difficult. To illustrate the expansiveness of these processes, in Arizona alone, there are two pending state adjudications for how to divvy up surface water from the Gila and Little Colorado rivers, the boundaries of which include more than half the state and flow through most of the tribal and federal land in Arizona. The Little Colorado adjudication includes almost 40,000 claimants, including the Navajo and Hopi tribes, in a judicial process that began in 1978, and has no end in sight. The quantity of water these rights represent is ostensibly vast, a key factor in the shifting water use and management landscape in western states.

The earlier sections outlined the complicated interplay between competing interests, on both institutional and practical levels that influence how the legal framework surrounding tribal reserved rights to water evolves, and the wide variation in its practical implementation (or lack thereof). From the literature, and historical record, we know factors like economic activity or nearby agriculture move the needle on settlement. Given these factors, what else is driving the shift from one state to the next in terms of institutions?

I analyze empirically which factors are correlated with a tribes’ decision to begin Winters proceedings, and the probability that parties will come to agreement and end negotiations. First I model a cross-sectional relationship between factors that influence the probability of a Winteres case starting. Why do some tribes enter into Winters proceedings and some do not, and do these factors influence whether a tribe will initiate proceedings earlier or later?

I then analyze which factors, changing over time, influence whether tribes begin to initiate Winters proceedings. I base this empirical analysis within the framework of discrete-time transition models, where a binary choice between staying in the current state, or switching to the next state is a function of past and present conditions, and the probability of switching in the current time period is conditional on “surviving” in the previous state up to that point.

As in latent regression analysis, a “switch” to the next phase implies that latent conditions have crossed a particular threshold. In the case of Winters, tribes will initiate a Winters proceeding if the benefits of obtaining enforced “wet” rights outweigh the costs of engaging in negotiations or settlements. For example, suppose $yit∈0,1$⁠, where $yit$ represents the decision to switch, and $yit=1$ for tribe i and time t if in time t that tribe decides to start negotiations. Suppose a latent variable, $yit*$⁠, represents the net present value of engaging in Winters proceedings, then $yit=1(yit*>0)$⁠.

For the first phase, the decision to begin Winters proceedings, at each time t, tribes choose to start negotiations or remain in the no-Winters state. They seek to maximize their expected utility, and face a cost of entering into negotiations, or potentially a lower growth trajectory without negotiations, which would mean they may remain at lower supplies of available water. Part of the decision process for tribes naturally is factoring in their beliefs about the probability of “winning” in negotiation, and how likely their chances are of plausibly receiving water rights in the long term. In essence, expected utility for this decision point is in part a function of the perceived probability of winning, which is a function of latent and observable tribal and state characteristics.

For the decision to end negotiations, parties also seek to maximize expected utility, and in each period t choose to either agree, or continue negotiating. Parties will agree if their expected future utility of agreeing is higher than the expected future utility of remaining in negotiations. Therefore, factors that influence costs and benefits of settling water and implementing the infrastructure necessary to use these water rights will impact how parties perceive these utilities.

I deploy a probit model specification to study which factors are associated with higher or lower odds of a tribal nation pursuing Winters enforcement procedures. These cross-sectional equations provide an initial look at whether both institutional/legal factors and environmental factors (including indicators of scarcity or competition for resources) may influence where we see Winters cases taken up and where we do not. Importantly, these are not causal estimates, but are a first pass investigating correlations.

Equation (1) represents the relationship between whether a tribe j institutes Winters proceedings at any point since 1908 (denoted by the binary variable “Winters” which takes on the value of 1 if the tribe engages in Winters and 0 otherwise) ^[32], and influencing political, environmental and demographic/economic factors. In this model I hope to capture factors that relate to the inherent balance of power between state, tribe and the federal government; whether there are entrenched on-the-ground de facto rights that have historically developed; inherent scarcity and competition for water resources; and controlling environmental factors:

where Disclaimer is an indicator variable taking on the value of 1 if the tribal nation’s surrounding state had original disclaimer clauses in their enabling act; BoRIrrigation is an indicator variable equalling 1 when there was at least one area of reserved Bureau of Reclamation irrigation land abutting a historic reservation (also in 1904); $Xc$ is a vector of socio-economic data, including real per capital income in the surrounding county (averaged between 1975 and 1984) and county population density on average between 1975 and 1984 ^[33]; $Ej$ includes a vector of environmental control factors per reservation, such as average streamflow (between 1975 and 1984), average daily temperature (1975–1984), average of daily precipitation (1975–1984), and the average Palmer Drought Severity Index (1975–1984) ^[34]. Finally, $Sj$ includes reservation-level average readings from the Soil Productivity and Drainage Indexes (the latter of which is reported as an indicator variable denoting if irrigation is required based on soil characteristics) ^[35].

Figure 4 shows American Indian reservation land in 1904 juxtaposed against Bureau of Reclamation claimed land in 1904 (areas used for proximity calculations).

I also estimate a cross-sectional relationship between the above factors, in addition to “Winters” activity in other watershed zones, and time to initiation (for those Winters rights that a tribal nation does move to get quantified and enforced):

where $Z1$ represents the average number of starts in previous five-year periods pre-1975 in other subwatershed (“HUC4” areas) regions than the one the reservation-in-question is in. The other covariates are the same as above.

Table 2 shows the results of the probit model Equation (1), analyzing the relationship between political, socio-economic and environmental factors and the probability a tribe will enter into Winters proceedings (dependent variable).

As Table 2 shows, close proximity to BoR reserved irrigation land early in this history (1904) is associated with significant increases in the probability of a tribe engaging in Winters proceedings down the line. Being in a “disclaimer” state is also positively associated with engaging in Winters.

Both of these findings make sense given the institutional context. In areas where, despite tribal reserved water rulings, the federal government facilitated early on the de facto development of water use directly off reservation, the more likely parties would have adjudicated conflicts over water. In other words, the on-the-ground staking out of water use and rights is associated with legal and institutional change after the fact. Being in a disclaimer state means technically for many decades jurisdiction between state and tribe was legally kept separate, perhaps engendering a climate where tribes may have thought their claims would have been more protected from state intrusion. This needs further investigation, especially mapping outcomes of Winters proceedings and settlement quantities, but the baseline correlations line up with the institutional premise.

Notably, other factors make intuitive sense as well. When there is more water in the system, evidenced by increased streamflow, increased participation on average, and a wetter climate (PDSI), there is an associated smaller likelihood parties will engage in Winters proceedings. In other words, when there is more water there is less to fight about. Interestingly, the taxonomic soil conditions are not significantly related to engaging in Winters debate, once other factors are controlled for.

Table 3 shows the results of modeling Equation (2), a cross-section of time-to-treat (i.e., how many years from 1908 does it take to start Winters proceedings?). The decision to enter into Winters is complex, and inherently dependent on availability of water, perceived balance of power, and perceived amenability of other parties to come to the table, settle, or get fair results in state courts.

Just like Table 2 showed that proximity to competing users (evidenced by BoR projects nearby, and surrounding county population density) and history of protection from intrusion (“disclaimer” state) was associated with increased likelihood of engaging in Winters proceedings, in this analysis these factors also are associated with earlier engagements with the process (fewer years from 1908). Notably, what could be perceived as increased competition for federal attention, funding, or protection, Winters starts in other watershed areas in previous years (pre 1975), was associated with a tribe initiating Winters proceedings later. Notably, the soil conditions do not significantly associate with later or earlier Winters starts.

Using the cross-sectional models, I analyzed latent factors that shifted a tribal nation’s probability of initiating Winters proceedings. These were based on both inherent and entrenched political forces, and also practical and environmental ones. Here, I model the decision to start or resolve Winters disputes by looking at time-varying factors, again, both political, economic, and environmental.

I model the decision to initiate Winters, and the decision to settle separately, each as functions of current and past individual- and state-level characteristics using a discrete-time hazard regression model. The dependent variable is the annual transition hazard rate, $hit$⁠, for tribe i, in year t, across state p. This represents the conditional probability of an event occurring at time t, given that no event has occurred up to time $t−1$⁠. To estimate this model, I use a pooled logit setup, where I allow for a separate intercept for each duration interval (year), but coefficients on regressors are constant over time (Cameron and Trivedi, 2005; Miranda and Trivedi, 2020) ^[36]. The following equation represents this general formulation for the conditional hazard model, where t is the time period in which tribe i chooses to negotiate or, in the resolution model, chooses to agree:

The $F(·)$ function is the cumulative density function of the logistic distribution. This generalization allows for variation across tribes ^[37] and time (⁠$xit$⁠), across tribes (⁠$zi$⁠), and across time (⁠$yt$⁠), where the discrete time changes are measured annually. The vector $(xit)$ represents tribal, time-varying characteristics, including climate and weather averages, and characteristics of the surrounding counties that had been included in the cross-sectional analysis. The vector $yt$ is a vector of time-varying characteristics, such as the presence of democratic presidential administrations in the 8-years leading up to time t, and policy shifts in funding that could influence the setting or potential outcomes of negotiations.

While this generalization, as mentioned above, does not directly allow for the constant to vary across time, I employ this framework to model the probability of transitioning to the next phase first for the decision to enter into negotiations, and then separately, to enter into the “resolution” phase. Many of the covariates are in both models, as the decision to engage in the Winters process is in part a function of the probability of succeeding in negotiation. Therefore, many coefficients are re-estimated for both base states, before negotiation and then during negotiation but before resolution.

Finally, I employ a pooled logit regression to model disparate years over time, across reservations, and use an indicator equal to one if a tribe chooses to transition in year t as the dependent variable. I estimate parameters by maximum likelihood. All errors are clustered to the state level.

In both models (decision to enter Winters and the resolution of rights), I model the transition to the new state as a function of environmental attributes such as climate, weather (annual patterns in precipitation and temperature), and average streamflow levels in the surrounding subwatershed regions (HUC4 areas); in addition to socio-economic factors such as surrounding county population density and real per capita income levels; with political indicators such as the presence of a democratic presidential administration in recent years (share of the previous 8 years democratic administration); federal government financial expenditures directed a state grants or projects; or outlays specifically for water resources; and entry and exit into Winters negotiations in other locations (a marker for competition amongst scarce federal resources, or a symbol of federal support and amenability in settling these cases).

For this analysis I exploit the observable nature of the negotiation process to study the relationship between changing state (and enforcing and defining rights) and institutions. Table 4 presents the results of this analysis. Column 1 shows results for the decision to enter into Winters proceedings, and Column  2 shows factors that correspond with the decision to exit.

Confirming earlier findings, increasing population density in the surrounding area is a strong positive correlate with the impetus to formally enforce and quantify Winters rights. Notably, increasing population density is not associated with negotiations coming to an end (which is intuitive considering it would imply more competition, and a potentially costlier adjudication). A run of democrat Presidential administrations is positively associated with starting Winters rights. Resolutions of Winters rights elsewhere, however is negatively correlated with starting proceedings, as are Winters starts elsewhere. This again could be an indicator of potential competition, and a signal to tribes to hold off until less-competitive times.

Current federal expenditure toward states are negatively correlated with starting Winters proceedings, while federal expenditures specifically for water are associated with higher likelihood of initiating proceedings.

For resolving Winters proceedings, resolutions elsewhere has a negative correlation with odds of settling. Federal outlays for water-specific projects are positively associated with resolutions, whereas general federal outlays in the current period to states are negatively associated.

Notably, the environmental factors like rainfall and drought are not significantly associated with starting or ending Winters proceedings. More research is warranted, however, to bring more granularity in analyzing the impact of these weather-and environmentally-related attributes. These are not causal estimates, but provide some of the first observable correlates linking political and legal institutions, with on-the-ground measurements of use, population pressure and environmental conditions.

The analysis presented in this paper outlines important institutional factors that underlie the trajectory of enforcing what are known as “Winters” rights. The original Winters Supreme Court case is often presented as the seminal case protecting water rights for American Indian tribes. Yet despite clear case law from the federal government, the practice of implementing and enforcing Winters rights has not been straightforward. Where Winters rights have been enforced and adjudicated is a function not only of the need for legal clarification, but shifting balance of powers between state, federal, and tribal interests. The federal government, throughout the twentieth century, has explicitly allowed states to either intrude, or not, into tribal and federal rights and jurisdictions. Understanding the fluctuations in the balance of power, and the relationship with on-the-ground water usage is pivotally important for understanding how these relationships evolve. Water management challenges still present major sources of conflict between interested parties. From an examination of history, controlling the legal right, or the de facto resources are important levers of power.

This paper presented initial evidence of the political influences that impact how and when Winters rights are enforced. The immediate years after the Winters ruling saw large investments in irrigation infrastructure, even for parties without superior or clean rights, and case law that reflected these shifting dynamics between state and federal players in the pursuit of developing the western frontier.

Both qualitative and empirical examinations of Winters implementation and conflict resolution highlight key findings. Both environmental conditions and political conditions matter. Competition for resources, with the system being already fully appropriated plays a key role in how and when conflicts resolve. Engaging with the Winters process to enforce and quantify rights requires involvement of both federal and state governments, and respective physical, political and economic capital. Econometric results indicate that the more tribes that initiate Winters proceedings elsewhere, the less likely a tribe is to initiate their own. On the flip side, when more Winters cases settle and get ratified by Congress, the odds of settlement elsewhere increase too (perhaps due to impatience over potential resource availability, or perhaps due to amenability of Congress to ratify, or both).

Empirical results also show indications that on-the-ground complexity, competition, and entrenchment also matter. The longer a negotiation plays out, the less likely the case is to being settled at all.

Results also support the notion that monetary investment on the part of the federal government matters. The odds of initiating negotiations and settling increase significantly when the federal government invests in water resources. This is a potentially powerful opportunity for the federal government, who has long taken a soft hand in helping to settle these outstanding conflicts, can make real impacts with credible funding and support.

From a big picture perspective, there is certainly still room for the federal government to help concretely to enforce reserved rights, and protect tribal sovereignty from intrusion of state power. As challenges in managing surface water persist, and become even more difficult as longstanding drought conditions continue to create scarcity in water resources, there are also indications the federal, state, tribal triad is finding innovative ways to contract and work together in managing water.

For example, in very recent years, the Jicarilla Apache Nation joined forces with the State of New Mexico and the non-profit philanthropic organization The Nature Conservancy to deploy settled Winters rights in new ways within the San Juan basin. The cohort came together to assert jurisdiction and authority over the federal government, finding non-traditional alliances between historically adversarial parties. What has resulted has been a cohesive and efficient arrangement to lease the Jicarilla nation’s water to the state for environmental flow purposes and species restoration. The agreement has produced positive results for the tribe, state and basin as a whole ^[38].

On the flip side, tribes still find themselves in a holding pattern created by over a century of government inaction and obfuscation. In June of 2023, Justice Neil Gorsuch, writing the dissenting opinion for the U.S. Supreme Court case Arizona v. Navajo, remarked that the Navajo Nation had a simple ask: “They want the United States to identify the water rights it holds for them […] And if the United States has misappropriated the Navajo’s water rights, the tribe asks it to formulate a plan to stop doing so prospectively (Gorsuch, 2023).”

Justice Gorsuch’s dissent was in response to the 5-4 opinion of the court to reject the United State’s obligation to identify and account for Navajo Nation water rights in the Colorado River, despite the clear de facto and de  jure understanding that the United States had legally granted such rights in establishing the tribe’s reservation in 1868. In his dissent, Justice Gorsuch compared the experience of the Navajo Nation to that of an annoyed consumer: “To date, their efforts to find out what water rights the United States holds for them have produced an experience familiar to any American who has spent time at the Department of Motor Vehicles […] The Navajo have waited patiently for someone, anyone, to help them, only to be told (repeatedly) that they have been standing in the wrong line and must try another.”

“Everyone agree[d],” Gorsuch continued, that the tribe in fact received enforceable treaty rights; that the U.S. currently holds a portion of them in reserve, but they have never been assessed. The majority opined they need not be. From Gorsuch’s perspective, the Court’s opinion makes clear that the “government’s constant refrain is that the Navajo can have all they ask for; they just need to go somewhere else and do something else first (Gorsuch, 2023).”

Like the Fort Belknap Indian Community, Navajo Nation is a case in point for a situation many tribes find themselves in. Situated in the northern Arizona desert, nearly half of the tribe’s reservation is bordered by streams, rivers or reservoirs that branch off of the Colorado River. Yet its community has very little usable water to show for it, and the tribe’s rights to the Colorado River still have not been adjudicated, despite its legal efforts to do so since the 1950’s. Approximately 30% – 40% of households on the Navajo Nation do not have piped water (Tanana, 2021), and the average person on Navajo’s reservation uses just 7 gallons of water per day, less than one tenth of the average amount for the typical American elsewhere (Gorsuch, 2023).

These statistics are similar elsewhere in Indian country. Approximately 58 out of 1,000 Native American households do not have access to indoor plumbing ^[39]; and nearly 30% of homes surveyed by the Indian Health Service (IHS) needed improvements in sanitation for sewer and/or solid waste systems ^[40]. Another recent study found that compared to white households, Native American households are 19 times more likely to live in a home without indoor plumbing (Tanana et al., 2021). Households without plumbing or piped water often rely on hauling-in water, which is often several magnitudes higher in cost per acre foot to procure. Part of this is due to the fact that the federal government failed to provide infrastructure to divert supplies to tribal nations, exacerbated by the lack of quantification and funding of tribal water rights. In 1910, for example, shortly after the Winters ruling, the U.S. Commissioner of Indian Affairs stated that there were millions of acres of irrigable land on Indian reservations ^[41], ostensibly ready to be watered, yet by the 1970’s, only about 7 percent of such acreage had been irrigated (Wolfe, 1992).

Some of the tide is turning, especially as the federal government has increasingly used their role to help in resource management. In August of 2021, for the first time in its 99-year record-keeping history, the US Bureau of Reclamation declared an official shortage of water in the Colorado River ^[42]. While the western United States has been plagued by historic drought conditions for the last two decades, adding to the confluence of weather and climactic factors are long-standing conflicts over surface water rights and allocations from the Colorado River. The waterway’s fundamental framework for sharing its water amongst its seven basin states (four upper and three lower) and Mexico has been largely unchanged since it was written in 1922, despite drastically changing conditions in the water supply and population pressures. Meanwhile, the institutions that developed to define and enforce these rights have incentivized over-use and mitigated key opportunities for water users to trade in order to reallocate the resource in times of scarcity. Large-scale conflicts over ostensibly vast quantities of water have become entrenched in the system and culture of water use in the west.

In May of 2023, nearly two years after the 2021 shortage decree, and 101 years after their initial compact, representatives of the three lower-basin states drawing from Colorado River water finally made a deal to voluntarily conserve and reallocate. This came directly as a result of the federal government stepping in and threatening to impose mandatory, and more severe, cuts to water usage across the board. If enacted, these cuts would have represented the first time that the US federal government would have intervened to allocate Colorado River water to states ^[43]. Under the states’ recent compromise, California, Arizona, and Nevada agreed to reduce water use by at least 3 million acre-feet by 2026 in exchange for $1.2 billion in compensation from the federal government. The combination of credible threats by the government and the promise of funding for infrastructure proved to be the most powerful tactic to induce conservation and reallocate on a broad scale ^[44].

This study presents initial empirical findings highlighting key institutional and historical contexts. Future research can add granularity, detail from tribal nations’ economic activities, land use, and governance, in addition to modeling more explicit outcomes of Winters cases. There are tremendous opportunities for more effective and equitable resource management in terms of water in the western United States. The federal government, scarcity and competition all have played important roles in water management, and will continue to do so in the intervening years ahead.

Special thanks to NBER and the Resnick Sustainability Institute at Caltech for generous support.