Abstract
Federal land management has always been contentious, but recent years have seen a rise in legislative activity demanding transfer of federal land to state control. Can states afford to assume land management without increasing taxes or other revenue enhancement actions? We find that (1) on average, federal lands are not likely to be as economically productive as private lands, (2) states are likely to have management costs equivalent to federal agencies, and (3) states can cover land management costs with land-based revenues if they have access to fossil fuels and timber resources, and prices for these commodities are relatively high. (JEL R14, R52)
I. INTRODUCTION
Ownership of the public domain has long been a contentious issue between western U.S. states and the federal government. The root cause of the long-simmering battle is public land policy in the West; the recent armed confrontations between private citizens and federal authorities in Nevada (Nagourney 2014) and Oregon (Johnson and Healy 2016) are simply the latest manifestation of over 100 years of western frustration. The legislatures of Arizona, Colorado, Idaho, Montana, Nevada, Utah, and Wyoming have all passed or debated legislative actions aimed at transferring federally administered public lands to the states (Keiter and Ruple 2014). For example, Utah’s HB 148 demanded the transfer of over 31 million acres from federal to state control (exempting national parks, wilderness areas, and military reservations), while Arizona’s HB 2321 asserted claim to all federal land within the state’s borders.
Advocates of state control advance three economic arguments: first, federal land management policies give too much weight to nonmarket services and too little weight to market commodities. They believe that states can better manage public land to produce market commodities, generating more jobs and income in rural regions. Second, bureaucratic and regulatory inefficiencies make federal land management agencies less cost-effective than state agencies, which can manage land at a much lower per-acre cost. Third, enhanced revenues and improved cost efficiencies mean that states can assume management responsibilities with no need for additional revenues through increased taxes, land sales, or other measures.
In this study we examine each of these arguments in turn. A short history of federal land management policies provides the necessary context for our analysis of land ownership and quality. We use county-level geographic information system (GIS) data for eight states of the Mountain West to empirically model differences in land quality by ownership, where the model is specified by the factors identified by historians as important to allocation of land. We find that a selection effect governs land ownership in Mountain West states: in short, private lands are concentrated in counties with greater agricultural, energy, and timber potential, while federal lands are concentrated in topographically rugged counties that have less arable land and energy resources. This raises the obvious question: if states gained title to less economically productive federal land, would they be able to generate revenues sufficient to cover land management costs, as claimed?
We use a case study of land management in Utah to examine the fiscal assumptions made by advocates of a large land transfer. Analysis of the programmatic budgets of the major land agencies operating in Utah finds that, after accounting for cross-agency subsidies, state and federal agencies have similar per-acre management costs. Finally, the key to fiscal affordability in Utah and many other western states is revenues from fossil fuel extraction, which far exceed revenues associated with extraction of other minerals, grazing, timber, or recreation. Our simulation model reveals the market, production, and royalty conditions under which the state can (or cannot) generate land-based revenues sufficient to cover management costs. While the fiscal analysis is unique to the state of Utah, the national-level budgets for major federal land management agencies suggest that the issues raised in this study are applicable to many other public land states.
II. WHO GOT WHAT: LAND OWNERSHIP AND QUALITY IN THE MOUNTAIN WEST
Why So Much Public Land in the West?
For much of America’s history, federal land policies have been factious; however, they were crucial in establishing a strong, centralized federal government and encouraging settlement of vast new regions acquired through purchase or treaty. Beginning with the Land Ordinance of 1785 during the Articles of Confederation, early U.S. land policy was focused on using land ceded from the original states to pay down the Revolutionary War debt, to compensate military veterans, and to fund ongoing government activities in an uncertain tax environment (Gates 1968, 61). Later land policies generally disposed of lands at nominal prices, and numerous laws were enacted to sell, grant, or otherwise transfer federal lands to promote and accelerate settlement of the Middle and Far West (e.g., the Pre-emption Act of 1841 and the various Homestead Acts). Since 1781 some 816 million acres have been transferred to private ownership. Most of the land transfers—97%—occurred prior to 1940 (Gorte et al. 2012).
Other federal policies provided for land withdrawals—lands to be managed for particular public purposes. Provisions for state trust lands were included in the enabling acts of new states; as land was surveyed, these reserved sections were granted to the new state without further action (Walker 2006). Congress allowed states to select in lieu lands from elsewhere in the public domain when their reserved lands were already occupied by homesteaders or dedicated to other purposes (Culp, Conradi, and Tuell 2005). Trust lands were then managed or sold by states to fund designated beneficiaries such as public schools. Some 471 million acres have been conveyed to the states, primarily in the form of trust land grants. Other withdrawals for specific public purposes included military fortifications, mineral reservations, and conservation.
Large land grants to railroads notwithstanding, U.S. land disposal laws were heavily influenced by the Jeffersonian ideal of aiding the small yeoman farmer.1 The Pre-emption Act of 1841 and the Homestead Act of 1862 both granted parcels of up to 160 acres in return for a nominal fee and, in the case of the Homestead Act, modest land improvements believed to indicate that the recipient was a bona fide settler intending to remain on the land. The 160 acre limit worked well with the ample precipitation and rich soils of the Midwest, but conditions changed west of roughly the 98th to 100th meridian. Here, soils, aridity, or topography, or all three, made homestead success an uncertain prospect. However, the abundant land and sparse population of the Far West during the 1850–1870 period meant that physiographic constraints were not initially binding on western settlers.
The vast majority of land in the Far West was not compatible with small-scale agriculture and was, instead, better-suited as rangeland. However, a single cow-calf unit might require 25 acres or more to support, meaning that the acreage limit of the Homestead Act would condemn settlers to a subeconomic operation. Libecap (2007) notes that in some portions of the West upward of 10,000 acres may have been required to achieve the appropriate economy of scale. Gates (1968, 466) found that many early arrivals in the Far West had, by the 1870s, cobbled together very large landholdings using a combination of actions, some legal and some not. A cattleman might have one or more of his ranch hands apply for land along a stream, with the water being put to beneficial use on the larger operation. Then each hand, having patented a homestead in his name, sold it to the original cattleman. This achieved two goals. First, the ranch was larger and thus getting closer to the scale needed to survive as an economic operation. Second, having senior rights to scarce water meant that a comparatively small, established ranch could effectively control access to thousands of acres of the public domain.
By the late 1870s, the public domain was perceived as becoming scarce, with the frontier famously declared “closed” in 1890 (Turner 1893, 1). In addition to achieving economies of scale in ranching, senior water rights held by large ranches might leave little water available for newly arriving settlers, who then had to look elsewhere. Another common means of preventing entry by new settlers was to graze open rangelands down to the ground, making the land unattractive as a homestead (Libecap 1989, 61). Further, the absence of a functioning legal system and an influx of “land pirates” to the increasingly populated Far West were key reasons for the development of claim clubs (Murtazashvili 2013, 66). Whereas clubs were initially developed to enforce property rights among their members, many ranchers also found such clubs to be useful in dissuading new settlers from establishing homesteads. First, claim clubs often illegally transformed open range into regulated common property by posting notice that the open range in a given region was at capacity and that new entrants would not be tolerated. Second, club members colluded to limit bidding at public land auctions, threatening those who would bid against the club with “physical harm, if not death” (Murtazashvili 2013, 148).2
The latter part of the nineteenth century saw successful passage of the Mining Act of 1872, a legislative act written for (and by) western miners in accordance with the conditions of mining in a western landscape. Fossil fuel resources remained immune to comprehensive legislation until the Mineral Leasing Act of 1920 declared fossil fuels the property of the federal government; until then subsurface rights were governed by terms of the particular law under which settlers filed their land claims.3 Despite the advocacy of John Wesley Powell and General Land Office Commissioner Samuel S. Burdett in favor of homesteads of up to 2,560 acres, ranchers and farmers were never able to achieve such advantageous legislation as miners had (Gates 1968, 419; Libecap 1989, 63). If anything, the perceived scarcity of land in the West and the strong commitment by Congress to the small farmer resulted in new land disposal legislation that continued to favor small-scale agriculture (Murtazashvili 2013, 211–212). The Desert Lands Act (1877), the Enlarged Homestead Act (1909, with a 320 acre maximum), and the Stock Raising Act (1916, with a 640 acre maximum) were all aimed at the small farmer; the Reclamation Act (1902) was designed to provide irrigation water to aid further disposal of western lands to small farmers (Wilkinson 1992). Each of these acts initiated a rush of new land claims by farmers in the Far West, some of whom were successful and many of whom were not. Gates (1968, 505) found that of 89,000 land entries in Wyoming between 1910 and 1934, less than half were patented.
From 1891 through the middle of the twentieth century, emphasis gradually shifted from land disposal to retention and management of the remaining federal lands. In addition to new legislation designed to encourage disposal (outlined above), other presidential and congressional actions of this era focused on land preservation and conservation, eventually leading to establishment of the National Park System, the National Forest System, and the National Wildlife Refuge System (Alexander and Gorte 2007). The effective end of the land disposal era in 1940 left much of the western United States under federal ownership. The failure of land laws to recognize the realities of farming and ranching in rugged and arid regions, the actions of settlers in preventing new homestead claims, and the conservation movement of the late nineteenth and early twentieth centuries all resulted in the large federal landholdings of the west. In the eight mountain states, some 263 million acres—roughly 48% of the region’s total land area—remained under federal management in 2010 (Table 1; Figure 1).
Public Land Ownership in the Intermountain States
Federal Acreage by Land Management Agencies, 2010
Unrest in the West
This high concentration of federal lands has never set well with many in the West, especially politicians and rural residents. States began agitating for more aggressive transfer of lands to private hands in the 1910s, culminating in President Hoover’s 1931 offer to relinquish surface ownership but retain mineral rights in federal hands, an offer that was opposed on conservation grounds by eastern states and summarily rejected as inadequate by the public land states.4 Access to federal rangelands came to be governed by the U.S. Grazing Service under the Taylor Grazing Act of 1934. Tensions between states and federal regulators simmered through the 1940s and early 1950s as western congressional members in both the Senate and House worked to achieve land use and land disposal goals through coordinated efforts to manipulate the Grazing Service budget (Gates 1968, ch. 21; Peffer 1951, chs. 15 and 16).
The 1960s and 1970s marked a transition into the modern era of environmental management, with the enactment of federal statutes protecting air, water, habitat, and wildlife resources. In the realm of public lands management, Congress passed the Multiple-Use Act of 1960 and the Wilderness Act of 1964, which placed 9.1 million acres of U.S. Forest Service (USFS) land under restrictive land use rules. These acts explicitly recognized recreation and preservation as legitimate uses of public land alongside traditional mining, logging, agriculture, and ranching. Although no Bureau of Land Management (BLM) land had been designated as wilderness, the BLM began adding wildlife and recreation representatives to its advisory boards.5 In 1976 the Federal Land Policy and Management Act (FLPMA) expressly declared that the federal government would retain ownership of the public domain unless the Secretary of the Interior determined that disposal serves the national interest. The FLPMA also declared that public lands administered by the BLM were to be managed for multiple uses and values, and repealed almost 2,000 statutes addressing land disposal policies enacted in earlier decades (Skillen 2009).
By the end of the 1970s, stockmen, farmers, loggers, and miners of the West—the traditional users of public land—felt besieged. Representation of livestock interests on the BLM’s National Advisory Board had fallen from 87% in 1949 to 28% in 1976 (Cawley 1993, 75). From 1964 through 1980, active BLM grazing permits (as measured by animal unit months) had fallen by 27%, while during this same time designated wilderness in the contiguous United States had grown by over 150% (BLM various years). Two oil embargoes during the 1970s revealed how exposed the U.S. economy was to foreign sources of fuel and nonfuel minerals, and newly enacted environmental regulations were seen by many westerners as hampering development of these resources. Finally, in the late 1970s the Carter administration’s Roadless Area Review and Evaluation threatened even more wilderness designations in the West (this time on BLM land), and its “hit list” of western water projects had generated much controversy (Cawley 1993, 82).
All of these issues were significant factors in the “Sagebrush Rebellion” of the late 1970s and early 1980s. The West was looking at a future in which federal land use policies left little local influence as to how those lands would be used. The Sagebrush Rebels initiated a flurry of federal, state, and local legislation, court challenges, and demands for changes in federal land administration, all of which, for the most part, proved futile. The rebellion has never truly ended. In addition to the numerous Congresses since 1980 that have considered bills aimed at reducing federal land ownership, either through land transfers or by capping federal land acquisitions, the 1990s saw the Wise Use and County Supremacy movements geared toward greater local control over public land use (Alexander and Gorte 2007; Grossman and Bryner 2012, 9). The land transfer proposals and armed confrontations of the early twenty-first century represent yet another upwelling of western land policy concerns that, after 100 years, have yet to be resolved.
Using History to Specify a Current-Day Land Ownership Model in the Mountain West
Our historical review of public land disposal raises the question of land quality. If settlers controlled the bulk of land that was best suited for agriculture and ranching, and if the race to control subsurface energy resources resulted in an uneven distribution of fossil fuel rights, was the land left in federal ownership by 1940 less able to generate market commodities that homesteaders and other potential landowners would have required? Casting this question within the context of the current land transfer movement, what is the potential of the federal lands targeted by transfer advocates to generate market revenues needed to manage the land?
The bulk of federal land targeted for transfer is currently managed for multiple use by federal land management agencies. That is, with the exception of Arizona’s HB 2321, land transfer bills exempt national parks, wilderness areas, and military reservations. Hence, our modeling effort focuses on public lands that are currently managed for multiple use. The vast majority of this land is managed by the BLM and the USFS, with smaller portions administered by the U.S. Fish and Wildlife Service and by the National Park Service (e.g., Glen Canyon National Recreation Area).
Our county-level model explaining western land ownership is restricted to current-era information available in GIS databases (in our case, about 2005). The U.S. Geological Survey GAP status codes from the Protected Areas Database of the United States, known as PADUS (U.S. Geological Survey 2012), was used to define land ownership and management classes in 8,100 m2 pixels. Given that the land management flashpoint is publicly owned multiple-use lands, GAP status codes 3 and 4 for public land were of greatest interest. These codes identify lands that are subject to uses of “broad, low-intensity type (e.g., logging, OHV recreation) or localized intense type (e.g., mining),” or lands on which there are “no known public or private . . . restrictions” that prevent conversion to a different land cover. We define “multiple-use” land as land that can be used for several purposes with few (if any) restrictions governing land cover change.
PADUS includes private land only if it is under some form of development restriction, such as land owned by the Nature Conservancy. Private land is therefore defined as the complement of all land area classified as public land (or private land under restrictions) in PADUS.6 This assumption was cross-checked against the SAGEMAP database,7 which indicated that over 97% of lands not included in the PADUS database are private lands. The private, federal, and state multiple-use land classes were then converted to a percentage of total land area in a county, with large bodies of water such as Lake Mead, Lake Powell, and the Great Salt Lake eliminated from the calculation. Table 2 shows land ownership in the eight mountain states, as measured by private, federal, and state multiple-use land classes described above (plus remaining land in the “other” categories).
Percentage of Land Type, by State
Generally speaking, historians have asserted that land settled by private landowners in the West was well suited for particular purposes: either the land was good for production of crops, animals, or timber, or the land had known mineral value. Our historical review makes clear that the most arable land had been claimed by early settlers. Water availability and climate (growing season) are measured by the innate ability of land to produce useful chemical energy, independent of irrigation. An important measure is net primary production, the net addition to biomass per unit of time, given as grams of carbon per square meter per year. Settlers relying upon homegrown crops would have been attracted to areas with relatively high net primary production; our variable is the average net primary production of all pixels within a county. A measure of soil quality was based on soil capability class and captures the percentage of a county with poor-quality, non-irrigable soil (U.S. Department of Agriculture STATSGO2 database8).
Settlers generally tilled the valleys and avoided mountainous regions. Flat valleys were favored by early settlers because smallscale irrigation was possible even if a farmstead were located at some distance from a water source. A topographic index characterizing irrigation potential was developed by calculating the flattened area of each pixel. Pixels that exhibit large changes in elevation have larger flattened areas than pixels with little change in elevation. The ratio of the flattened pixel area to the original area was calculated for all pixels in the county, with the mean value of all pixel ratios in a county used as a measure of topographic variability. All else equal, settlers would have been attracted to regions with flat land, that is, lower index values.
Coal resources were important to settlers because fuelwood was scarce. Further, the burgeoning network of western railroads required both coal and oil, making lands with the potential for energy development highly sought by railroads as well. As a result, “towns sprouted around western coal deposits” (Zizzamia 2015, 193). Rumors of oil and coal often led to homestead or mineral claims on unappropriated land where deposits were suspected. For example, even after all of the remaining public domain believed to hold oil in Wyoming had been reserved in 1909, intensive exploration efforts in 1917 led to a situation where “every acre of land that was not withdrawn and showed reasonable promise of oil was leased or bought” (Ise 1927, 96). Our measure of energy potential is the percentage of land in a county overlaying a coal or oil basin. The sign of this variable will depend on the relative ability of private and federal agents to identify and claim suspected energy deposits under the shifting laws governing energy resources in the late nineteenth and early twentieth centuries. As such, we have no expectation on the sign.
Another crucial element in western land ownership patterns is the location of frontierera U.S. Army forts in the western territories. Military posts, necessarily sited at locations with cultivable land and water, were linked by Army-built wagon roads that also supported emigration and trade (Tate 1999; Miller 1979). Forts provided protection and provisions to early settlers along the various emigration trails and to workers building the transcontinental telegraph lines and railways, as well as being an often-used source of resolution in disputes among settlers. Homesteaders, in turn, supplied soldiers with agricultural products and fuel. We measured the distance from the center of a county to the nearest frontier-era (1850–1890) military outpost (Distance to fort). We expect to see more private land in counties that are closer to the location of forts.
As a means of subsidizing the construction of transcontinental railways, a railroad company could receive specified sections of federal land in a corridor adjacent to the railway and then sell those sections to private citizens. In many cases the land was of poor quality and railways could not sell the land (which then remained under federal control), whereas in other regions the railways had little trouble with sales (Henry 1945). We based a dummy variable on the route of the land grant transcontinental railroads: if a land grant railroad ran through a county, the variable took the value of one, and zero if not. We expect the best land to have been sold or retained by railroads in private ownership, with the least productive land returning to the public domain.
Finally, the Enabling Act for each state required the federal government to relinquish title to the state of specific sections of land (state trust lands). Arizona, New Mexico, and Utah received trust lands totaling four square miles, equal to sections 2, 16, 32, and 36 in each 36-section township; the other five states received only two sections per township (sections 16 and 36). If a specific section had already been appropriated by, say, federal withdrawal or a homesteader, the state and the federal government negotiated for an equivalent in lieu parcel of land elsewhere in the state. The Enabling Act variable took the value of one for Arizona, New Mexico, and Utah, and zero for the other states. Counties in Arizona, New Mexico, and Utah should have less private and federal land (and more state land) than counties in the other five states.
Table 3 reports descriptive statistics for the dependent and independent variables in our models. Private land and federal and state multiple-use lands account for an average of almost 83% of all land in the 276-county region. Much of the region (88%) consists of soils not suitable for irrigation due to soil quality or slope. On average, land in the region produces nearly 2,800 g of carbon per square meter per year, but some lands enjoy a microclimate enabling them to produce large quantities of vegetation. The average topographic index shows that most counties are relatively flat, on average, but other counties are quite mountainous. Further, an average county has about 32% of its land overlaying an energy basin, but this measure shows a high degree of variation across the region. Finally, some 28% of counties are located in states receiving four sections of land per township at statehood, whereas the remaining counties are located in states that were allocated only two sections.
Variables Used in Land Ownership Analysis (276 Counties)
An Empirical Model of Current Private, Federal, and State Land Ownership
The models for land ownership demonstrate a statistical relationship that is consistent with the arguments of historians (Table 4). Turning first to the model of the proportion of land in a county owned privately (column 2), one sees that as the percentage of soil in the county classified as nonirrigable increases, the proportion of private land falls. As the net primary production of land increases, the percentage of private ownership in a county increases. As land in a county becomes more topographically variable, less land is owned privately. Counties with a greater proportion of their land overlaying potential coal, gas, or oil resources have a greater percentage of private land ownership. In cases where the center of a county becomes more distant from frontier-era military outposts, less land in that county is privately owned. Land grant railroads running through a county are not a statistically significant factor in explaining the proportion of private land in a county. Finally, counties located in states receiving four sections of land per township have smaller proportions of land in private ownership. All of the variables except for land grant railroads are highly significant (p ≤ 0.01), and the model explains nearly half of the variation in the dependent variable. Thus, we find that today’s private land ownership echoes the factors identified by historians as influencing private land claims by settlers and early arrivals to western states. Further, the models suggest that private lands are likely to have greater than average agricultural, timber, and energy potential.
Land Ownership Models (276 Counties)
We find just the opposite for federal multiple-use lands (Table 4, column 3). As the percentage of a county’s soil classified as non-irrigable increases, more land is left in federal multiple-use management. As the average net primary production of a county’s land increases, a smaller proportion of land is found in the federal estate. The more mountainous a county, the more of that county’s land stayed with the federal government. As the energy potential of a county increased, the federal government was left with less land in that county. All else equal, the federal government has been left with land that is located farther from frontier-era forts. A land grant railroad running through a county results in less federal land, all else equal. Finally, the federal government holds a smaller proportion of land in counties located in states receiving four sections per township relative to those receiving two sections of land per township. Again, every variable in the model except land grant railroads (p = 0.05) is highly significant (p ≤ 0.01). On average, the federal government land holdings are larger in counties with less potential for agricultural, timber, or energy production.
State trust lands were initially allocated in what is best described as a random process (sections 16 and 36, plus sections 2 and 32 in Arizona, New Mexico, and Utah) regardless of the section’s characteristics, although the state could negotiate for in lieu sections of land elsewhere if a designated section was already appropriated. Consequently, our model of state multiple-use lands is not quite as strong statistically as for private and federal lands (Table 4, column 4). State land-holdings in a county are positively related to net primary productivity and the number of sections it received under the Enabling Act, and negatively related to steeper average topography and land grant railroads (p < 0.09). All other factors (nonirrigable soil, distance to forts, and energy potential) are not statistically significant. Given the relatively random nature of lands assigned to the state—and the fact that states sold much of their more valuable land to private ownership to build school trust funds following statehood—it is not surprising to find a relatively weak relationship between state ownership and the independent variables.
The models in Table 4 imply that private lands are concentrated in counties that are more amenable to agricultural, timber, and energy production. Federal lands, in contrast, are concentrated in counties that are more rugged and have less arable land, a greater proportion of poor, nonirrigable soils, and less energy potential. If a transfer of federal land were to occur, states may acquire land that is, on average, more limited in its ability to generate revenue than privately owned land, or even much of the land that states currently manage.
III. ARE STATES MORE COSTEFFECTIVE LAND MANAGERS?
Advocates of a land transfer argue that state land management agencies are more cost-effective than federal agencies laboring under more difficult regulatory and bureaucratic constraints. We address this question by examining the programmatic budgets of agencies operating in Utah, a large public lands state, tracing the flow of federal and state funds across various management functions to determine what agencies are managing which land management functions. The dominant federal land management agencies in Utah are the BLM and the USFS, which manage 22.8 million acres and 8.2 million acres, respectively. These agencies, along with the National Park Service (2.1 million acres) and the Department of Defense (1.8 million acres), collectively manage nearly all of Utah’s 35.0 million acres of federal land, or about 64.5% of Utah’s total land area.9 The School and Institutional Trust Lands Administration (SITLA, 3.4 million acres) and the Department of Natural Resources (DNR, 2.0 million acres) manage all but 2,700 acres of Utah’s 5.4 million acres of state land. Within the DNR, the primary division with land management authority is the Forestry, Fire, and State Lands (FFSL), which controls 75% (1.5 million acres) of DNR land.
Average revenues and costs over a five-year time period (2008–2012) show wide variation across BLM, USFS, SITLA, and FFSL agencies (Table 5, top). A simple interpretation of the numbers suggests that the BLM and SITLA are marvelously effective land managers—with per-acre profits of $8.46 and $24.25, respectively—whereas the USFS and FFSL lose money on their lands ($10.60 per acre and $5.95 per acre, respectively). In fact, one must examine the budgets more closely to fully understand the relationship of revenues and costs across public land management agencies.
Comparison of Revenues and Costs for Major Federal and State Land Management Agencies Operating in Utah (Fiscal Years 2008–2012 Average, Constant 2013 Dollars)
First, the USFS is responsible for many wildfire suppression activities on private, federal, and state lands and does not cross-bill other land management agencies for, say, the cost of tanker aircraft used to fight wildfires on BLM or state lands. Further, the BLM is responsible for managing oil, gas, and coal royalties on federal lands in Utah, so that all revenues for fossil fuel leases on USFS-administered land are counted as BLM revenues. The FFSL and federal agencies incur nearly all costs for wildfire suppression and mitigation on SITLA land, sparing the SITLA these expenses—and some of the FFSL’s postfire rehabilitation expenses are paid by the federal government. As another example, the SITLA does not incur the full cost of rangeland improvements or determining appropriate stocking levels on its range, instead relying upon the Utah Department of Agriculture and Food to cover salaries of range specialists and much of the infrastructure costs. In the case of isolated sections of SITLA land within the larger federal estate, the SITLA incurs little cost because it usually follows the stocking recommendations for nearby federal allotments.
Though still subject to error because we could not track all subsidies and cross-billing across federal and state agencies, a better measure would combine agency revenues and costs and divide by total acres managed (Table 5, bottom). If one combines the BLM and USFS budgets, per-acre revenues are $10.49, whereas per-acre costs are $7.06. Including average per-acre federal Payments in Lieu of Taxes (PILT) and Secure Rural Schools (SRS) transfers as costs of federal land management, per-acre costs increase by $1.43 to $8.49.10 On the state side, a combined accounting of SITLA and FFSL budgets yields per-acre revenues of $23.38 and per-acre costs of $8.52. Here we see that federal and state land management costs are actually quite comparable.11
Some might not consider PILT and SRS payments a true cost of land management as they are designed as compensation for states and counties and not directly incurred as on-the-ground management costs. Net of PILT/ SRS payments, federal land management costs are approximately 17% lower than state costs.12 A possible explanation is that federal agencies have managed to achieve scale economies in their management of large tracts of land, which the state has not been able to achieve on its smaller land base. A transfer of land from federal to state authorities may permit the state to achieve similar economies. Further savings may be possible if, as the state maintains, legal costs fall as the bulk of state land management decisions become exempt from the provisions of the National Environmental Policy Act.13
IV. COVERING MANAGEMENT COSTS OF A FEDERAL LAND TRANSFER
Utah’s Transfer of Public Lands Act14 calls for the transfer of 31.3 million acres of federal lands to the state of Utah from the current 35.0 million acres.15 Utah’s land management responsibilities would increase nearly sevenfold, from 5.4 million acres to 36.7 million acres. As a proportion of state area, Utah would become the largest owner of state lands among the 50 states, managing 68% of its land area. The cost for managing the 31.3 million transferred acres is estimated to be between $280 and $283 million per year.16
Given the state’s constitutional requirement for a balanced budget, the transferred lands must generate at least $280 million in additional revenues to cover the newly assumed costs. If land with the greatest potential to generate market revenues is already in private ownership, and if state land agencies are no more cost-effective than federal agencies, what are the conditions under which states can afford to assume the relatively large costs of a land transfer? Analysis of federal and state land management budgets over the fiscal years 2009–2012 period found that 84% of federal revenues ($324.9 million) were generated by oil and gas fees/royalties, with another 8% contributed by coal. Clearly, the state could not rely upon grazing, recreation, rights-of-way, or other fees to raise the needed $280 million. Though Utah is among the smallest of energy producing states, it has proven reserves of 613 million barrels of crude oil, 7.8 trillion cubic feet of natural gas, 268 million barrels of natural gas liquids, and 14.9 billion tons of recoverable coal, most of which is located on the federal estate (Boden et al. 2013). This suggests that a state-level management approach that aggressively harvests in situ resources could generate the necessary funds.
A Simulation Model of Future Oil, Gas, and Coal Revenues
The key parameters governing future oil and gas revenues are the path of future prices, the number and productivity of existing wells, the rate at which new wells are drilled and their productivity, and the share of royalties between the federal and state governments. The oil and gas scenarios use two alternative price paths. The first path is the official “reference” price projections for oil and gas by the U.S. Energy Information Administration (EIA) in 2014, and which we call the High Price scenario. The EIA’s projections, adjusted to Utah wellhead prices, show prices increasing steadily over time (Figure 2). The EIA reference price projection incorporates growing use of alternative energy sources and recently adopted drilling technologies (e.g., horizontal drilling). We use the EIA price forecast from an assumed transfer date of 2017 through 2035 as the High Price scenario. Over this time period the expected mean Utah wellhead prices for oil and natural gas are $92 per barrel (range, $77 to $109) and $5.10 per thousand cubic feet (range, $3.60 to $6.60), respectively. As an alternative to the EIA price path, we developed a Low Price path that reflects greatly accelerated adoption of alternative energy, modern drilling techniques, and expanded output by other oil producing nations. For the 2017 through 2035 period, the Low Price scenario has expected mean oil and gas prices of $62 per barrel (range, $40 to $86) and $3.30 per thousand cubic feet (range, $3.00 to $3.60), respectively.
Historical Oil and Gas Prices, with Projections to 2037
The rate at which new wells are brought into production depends upon exploration activity, the percentage of wells that are dry or otherwise not economically viable, the prices of oil and gas, and regulatory restrictions. The forecast model assumes two drilling rates: the historical rate at which wells have been successfully developed at given oil and gas prices, and a 15% increase in that rate to reflect a more aggressive development posture the state would likely adopt. Utah seeks title to both surface and mineral rights immediately upon transfer, and its legal position is that it has the right to 100% of the revenues flowing from public lands. Some of our revenue projections assume the current equal share (50%–50%) across federal and state entities on existing wells and a 100% royalty share on all newly drilled wells; other scenarios assume that 100% of all royalties flow immediately to the state of Utah.
Revenue projections for oil and gas are predicated upon the forecast price paths appearing in Figure 2 and forecasts of production from existing and newly developed wells. At the time of the analysis Utah had 13,500 active oil and gas wells. Wells are classified as producing primarily oil or primarily gas, but, with the exception of coalbed methane wells, wells of either type usually produce both products. Thus, every well has its own production path, called a decline curve, for two products, except for coalbed methane wells, which have a single decline curve for gas.17 The decline curve relies upon three parameters. The first captures the initial production volume, while two shape parameters map the path of production over time. Active wells were sorted into groups to estimate 12 decline curves based on three geographic locations and the primary product produced (oil or gas), with a thirteenth decline curve estimated for coalbed methane wells.
For each existing well with a history of production, the parameters of the decline curve reflect both the actual production history of that particular well and the average values of those parameters among wells in the same group. The relative influence of an existing well’s particular history in estimates of the parameters of its decline curve vary directly with how long the well has been in production and inversely to the difference between the well-level parameter and its group mean. Wells were evaluated for profitability over the time frame of analysis: as output level or price changes caused a well to fall below the zero-profit threshold, the well was abandoned.
Exploration and development of new wells was tied to oil and gas prices. Higher prices over time spur development activity, while lower prices slow exploration and development of wells. Decline curve parameters for wells drilled in the future and not yet having a production history were drawn from the joint distribution of the three decline curve parameters (initial production and two shape parameters) of the group to which such wells belong. Quarterly production was estimated for the time period 2017 through 2035 on existing and new wells for each of the two price paths. The volume of production from existing wells declines over time, but total volume for all wells may increase or decrease depending on the rate at which new wells are drilled in each quarter.
Revenues associated with oil and gas production projections appear in Table 6. Each scenario is based on assumptions regarding prices (high or low price paths) and drilling rates (at the historical level or at an increased level), and royalty shares (equal shares on existing wells and 100% to the state on new wells, or 100% to the state on all wells). Under the High Price path the state almost immediately recovers the costs needed to assume management of the transferred land using oil and gas revenues alone, regardless of assumptions of royalty shares or drilling rates. Revenues are relatively insensitive to the drilling rate, but they are quite responsive to the immediate 100% royalty assumption. Under the Low Price path, the rate at which existing wells become unprofitable accelerates, and the rate at which new wells are drilled slows so that production volume falls over time. Oil and gas revenues alone are not sufficient to cover land management costs unless the immediate 100% royalty share is achieved (last column). Even then, revenues at the end of the simulation period fall below those needed to cover management costs.
Oil, Gas, and Coal Revenue Scenarios (Millions of Constant 2013 Dollars)
The other major source of funding is revenues from coal resources. The future of coal mining in Utah is dependent on a complex set of economic, geological, technical, and political factors. The Utah Geological Survey (UGS) provided three coal production scenarios for this study: Low, Middle and High. Details of the coal scenario assumptions are provided by Stambro et al. (2014, ch. 13), but in all three scenarios the UGS assumed that there would be steady depletion at existing mines. The Low scenario was based on decreasing demand for coal due to greenhouse gas regulation, retirement of aging coal-fired power plants, rapid conversion to natural gas, and difficulties in permitting new mines. The Middle scenario assumes steady domestic demand, then slowly decreasing, as no new coal-fired power plants are built and older plants are shut down or converted to natural gas. This is balanced by increasing export markets and some new mines opening. The High scenario assumes that coal demand increases due to the development of successful carbon-capture technology, leading to revitalized coal markets and increased mine permits. The revenue projections under the Low, Middle, and High scenarios for coal production are shown in Table 6. Coal is valued at the EIA’s 2014 reference case projected minemouth prices for Rocky Mountain bituminous coal, with 100% royalty share flowing to the state. Under the Low and Middle scenarios total coal revenues fall from $51 million in 2017 to $30 million (or less) by 2035. Only in the High scenario does coal revenue increase, rising from $51 million in 2017 to $77 million in 2035.
Utah’s ability to cover the additional $280 million cost of land management rests upon high commodity prices. If oil and gas prices are high, the state can cover its costs at current drilling rates while waiting for the 100% royalty share to become effective. Revenues from other sources, including coal, are not required. If oil and gas prices are low, then the state needs to assure an immediate increase in drilling rates and/or an immediate 100% royalty share on oil and gas. Given the slow exploration and development of new wells under the Low Price forecast, the state also must rely on high coal revenues to cover reduced oil and gas revenues. If oil, gas, and coal revenues are low, Utah will not be able to cover management costs from these sources. Instead, the state would have to develop revenue streams more aggressively from other productive activities taking place on public lands, such as recreation, grazing, and timber. However, these activities currently generate relatively small amounts of revenue; even doubling or tripling the amount of revenue they generate would not make up for low commodity prices and production.
V. CONCLUSIONS
Western states have become home to large amounts of federally managed land despite strong efforts to dispose of land to private ownership throughout the nineteenth and early twentieth centuries. Nearly all federal land transfers to private owners occurred prior to 1940, long before the official era of land retention began in 1976 with the Federal Lands Policy Management Act. The land retention and management policies of the twentieth century and, especially, the FLPMA in 1976, led directly to the Sagebrush Rebellion of that era, as well as the current calls by Mountain West states to transfer title of most of today’s federal estate to state authorities. Our statistical analysis found that the factors historians have identified as affecting the settlement of western land—namely, soil quality, aridity, topography, and energy potential— echo through to present-day patterns of land ownership in the Mountain West.
The land ownership models reveal that lands with the greatest potential to generate market revenues are, on average, already privately owned, with the remaining public domain having less commercial potential. That said, the vast federal estate remains home to valuable deposits of energy and timber, so relatively large revenue streams are possible with a large land transfer to states. Close examination of the programmatic budgets of federal and state land agencies, though, finds no cost advantage for state land management agencies over federal agencies. Consequently, management by state agencies is likely to incur costs similar to federal agencies; we estimate Utah’s cost to be about $280 million per year (constant 2013 dollars).
State officials have expressed a goal to manage public land using revenues generated by the land received in a transfer. The largest source of revenue, by far, is royalties and fees from extraction of oil, natural gas, and coal. Our modeling of fossil fuels production finds that Utah should generate revenues in excess of costs if fossil fuel prices are relatively high. If low prices prevail, then states would need to secure an immediate 100% royalty share rather than the current 50–50 split. Even then, low prices increase the rate at which existing wells become unprofitable and decrease the rate at which new wells are drilled. Both of these situations lead to declining revenues over time, eventually putting revenues below costs.
While our analysis is directly applicable only to states with fossil fuel resources, research by O’Laughlin (2014) in Idaho indicates a similar conclusion for states with substantial timber resources. O’Laughlin examined a transfer of 16.4 million acres of USFS and BLM land to the state of Idaho, a state that has few fossil fuel resources but abundant timber. Under a low stumpage price/ low output scenario, he reports the state would fall $111 million short of management costs, but under a high price/high output scenario, the state could gain up to $24 million per year.
Regardless of the resources harvested from public lands (fossil fuels or timber), state budgets would be tightly tied to commodity prices and production. Given that only one of the Mountain West states (Arizona) is legally permitted to carry a deficit across budget cycles (National Conference of State Legislatures 2010), this suggests that a land transfer could introduce volatility into the budgetary process. For example, in August 2013 Utah wellhead oil prices reached a high of $92.60 per barrel, well within the price range for the High Price scenario. Two years later, in August 2015, the price had fallen to $33.78 per barrel, below the lowest price anticipated in the Low Price scenario. Public lands states such as Utah are price takers in global energy markets so there is little a state or its producers could do to reduce price volatility.
Forced to absorb volatility in commodity prices, states could perhaps build a rainy day fund dedicated to public lands management. Galle and Stark (2012) review the rainy day fund literature and find that both voters and politicians focus on the present rather than the future (present-bias). The result is that state rainy day funds tend to be chronically underfunded. Another problem is that the economic conditions under which withdrawals from these funds are warranted tend to be very unclear and subject to political discretion. If states plan to use rainy day funds to minimize fiscal exposure to commodity price volatility, the size of the fund contributions and the terms of withdrawal must be carefully delineated. Another issue facing states planning to rely upon fossil fuel revenues to manage public land are future developments associated with climate change policies. Coal resources already suffer a cost disadvantage relative to natural gas; if climate change policy includes a carbon tax or permit system, then oil, gas, and coal would be put at a further disadvantage relative to alternative energy sources with fewer carbon emissions.
Finally, we note that if land-based revenues are insufficient to cover management costs, states may be tempted to sell land rather than retain it in state ownership. Indeed, amenity-based migration would suggest that nonmarket services, such as an outstanding viewshed or easy access to high quality outdoor recreation, could make some land very valuable. While some land may indeed be sold, two factors work against large-scale disposal to private ownership. First, the Enabling Act of every state requires the proceeds of land sales to be shared between the state and federal government, with the state receiving 5% of the proceeds and the remaining 95% returned to the U.S. Treasury. The only fiscal justification for the state to sell land would be if the present value of the expected flow of revenues, net of management costs, were less than 5% of the market value of the parcel. Further, if states do not collect a property tax (e.g., Utah), sales to private parties will generate only a one-time inflow of revenue. Second, any state effort to dispose of newly acquired public lands is likely to encounter fierce opposition, and not just from groups with an interest in continuing the flow of nonmarket values from public land. In writing about the first Sagebrush Rebellion, Green (2015) and Cawley (1993) report that when the Reagan Administration proposed large-scale sales of federal land under the 1982 Federal Real Property Initiative, it encountered strong opposition from ranchers, farmers, miners, and loggers, the very interest groups most frustrated with modern-era federal land policy. Libecap (2007) notes that such constituencies develop a stake in the status quo—in this case through the flow of rents associated with public land ownership—even if the status quo property rights allocation is economically inefficient. Large and unsettling uncertainties would accompany the radical change in property rights represented by a shift from public to private ownership. Libecap (2007, 259) writes, “Path dependencies in property rules are real, and they have dominated the economic history of resource use in the West.” A land transfer from one public entity to another is a more incremental change incurring smaller political transactions costs than large-scale land sales and less likely to disrupt the existing flow of rents, at least for commodity producers.
Acknowledgments
We thank Lassina Coulibaly and Ben Crabb for research assistance, and Flora Shrode for technical editing. We also thank Kathleen Clarke, John Harja, and Tony Rampton of the Utah Governor’s Public Lands Policy Coordination Office (PLPCO) helpful insights and discussion throughout this research project. The PLPCO funded this study. The views contained in this article do not necessarily reflect the views of the State of Utah or PLPCO; the authors are solely responsible for the research and conclusions contained herein.
Footnotes
The authors are, respectively, professor, Department of Applied Economics, Utah State University, Logan; Stambro Consulting Group, Hawthorn Woods, Illinois; senior research statistician, Kem C. Gardner Policy Institute, University of Utah, Salt Lake City; senior research analyst, Kem C. Gardner Policy Institute, University of Utah, Salt Lake City; research analyst, Kem C. Gardner Policy Institute, University of Utah, Salt Lake City; and professor, Department of Economics, Weber State University, Ogden, Utah.
↵1 Railroads patented 131 million acres of the 156 million acres contained in the original land grants (Henry 1945, 172).
↵2 Whereas mining claim clubs (mining districts) had a profound impact on western water law, mining claims were small and often abandoned and thus had limited influence on the allocation of the public lands. Timber interests also formed claim clubs, using techniques similar to those of ranching claim clubs to prevent entry by others. Libecap (1989, 53–60) documents the transactions costs of engaging in fraudulent land acquisition activities on timber lands.
↵3 Under some land disposal legislation both surface and subsurface rights transferred to private ownership, whereas under other legislation subsurface rights were retained by the federal government. Ise (1927, 291) comments that if all of the public domain had been “fit for cultivation . . . the federal government would never have owned any oil lands.”
↵4 See Peffer (1951, 205–9) and Gates (1968, 527–28) for a description of the Hoover proposal, which was introduced by a presidential letter stating, “Our Western States have long since passed from their swaddling clothes and are today more competent to manage these affairs.” Utah Governor George Dern replied, “The Western States appreciate the compliment of being assured that they are now man grown . . . but they can not [sic] help wondering why they should be deemed wise enough to administer the surface rights but not wise enough to administer the minerals contained in the public lands” (Cawley 1993, 73).
↵5 The Grazing Service and the General Land Office were combined in 1946 to form the BLM.
↵6 This measure of private land ownership does not account for 45.9 million acres of land where the surface is privately owned but subsurface mineral rights are retained in federal ownership (BLM 2014). Only a portion of split estate land will contain economically recoverable energy resources.
↵7 U.S. Geological Survey. SAGEMAP: A GIS Database for Sage-grouse and Shrubsteppe Management in the Intermountain West. Available at https://sagemap.wr.usgs.gov/ (accessed February 21, 2017).
↵8 Available at www.nrcs.usda.gov/wps/portal/nrcs/detail/soils/survey/?cid=nrcs142p2_053629 (accessed December 7, 2016).
↵9 For perspective, the federal government’s estate in Utah is a bit larger than the total land area of Florida.
↵10 PILT payments are used to compensate counties for lost tax revenues due to federal ownership of land, whereas SRS payments compensate rural counties for reduced timber revenues from USFS land.
↵11 Other reports (see, e.g., Fretwell and Regan 2015) often overlook the subsidies associated with federal land management practices (e.g., fire suppression) provided to state agencies. Another well-known study failed to include royalties flowing directly to Washington, D.C., which are later shared equally with the state generating the royalties (see Intertech Services Corporation 1994).
↵12 Some portion of overhead costs for federal expenditures is incurred in Washington, D.C. We were unable to allocate these costs to the Utah offices of the BLM and USFS.
↵13 State land management would still be subject to the provisions of the Clean Air Act, the Clean Water Act, the Endangered Species Act, and other broadly applicable federal regulations.
↵14 Utah Code Ann. 63L-6-101 to 104.
↵15 Some 3.7 million acres would remain under federal control in military reservations (1.8 million acres), national park units (0.9 million), and designated wilderness areas (1.0 million acres).
↵16 Details are provided by Stambro et al. (2014). The higher cost estimate is based on per-acre management costs (see Table 5). The lower estimate of $280 million acknowledges the cost advantages that may be achieved by state agencies based on experience of managing specific programmatic activities (e.g., managing fish hatcheries), and accounts for some changes in the flow of expenditures due to decreased federal acreage. The largest direct land management costs under the second approach are for wildfire ($87 million), rangelands ($84 million), and forests ($55 million).
↵17 A decline curve relates the expected production rate of a well to the elapsed time since the date the well first produced. Decline curves could follow an exponential, harmonic, or hyperbolic path.
