Volume 43, Issue 4 pp. 1571-1594
Featured Article
Full Access

The Economics of Groundwater Governance Institutions across the Globe

Eric C. Edwards

Corresponding Author

Eric C. Edwards

Correspondence may be sent to: E-mail: [email protected]Search for more papers by this author
Todd Guilfoos
First published: 27 September 2020
Citations: 14

Eric C. Edwards in the Department of Agricultural and Resource Economics, North Carolina State University, Raleigh, North Carolina, USA.

Todd Guilfoos in the Department of Environmental and Natural Resource Economics, University of Rhode Island, South Kingstown, Rhode Island, USA.

Editor in charge of manuscript: Daniel Petrolia

Abstract

This article provides an economic framework for understanding the emergence and purpose of groundwater governance across the globe. We examine ten basins located on six continents via an integrated assessment along three dimensions: characteristics of the groundwater resource; externality problems; and governance institutions. Groundwater governance addresses local externalities to balance the benefits of reducing common pool losses with the costs of doing so. While broad, basin-wide solutions to open access pumping are limited, spatially localized externality problems raise the benefits of management actions, allowing for the implementation of more stringent pumping controls in certain areas.

Introduction

Groundwater aquifers supply drinking water to approximately 50% of the world's population (Connor 2015) and provide 43% of the total water consumed in agricultural irrigation (Siebert et al. 2010). The depletion of the world's groundwater systems in the largest irrigating countries—India, China, and the United States—threatens the sustainability of the world's food supply (Aeschbach-Hertig and Gleeson 2012). Collateral effects of aquifer drawdown include problems such as land subsidence and seawater intrusion. Severe cases of subsidence are found in Mexico City, Bangkok, Shanghai, and California's Central Valley, permanently reducing water storage with measured declines in land elevation of tens of meters in some cases (Konikow and Kendy 2005). Seawater intrusion reduces the availability of freshwater to coastal populations and can render drinking and agricultural water supplies unfit for use, with severe cases in Oman and California (Zekri 2008; Barlow and Reichard 2010).

Despite the high costs of continued overexploitation, successful groundwater governance has been elusive in many regions (Aeschbach-Hertig and Gleeson 2012). The world's aquifers remain largely open-access resources, with overlying landowners free to pump water subject to few restrictions (Konikow and Kendy 2005). Efforts to understand the economic factors underlying groundwater governance, including where it emerges, its form and function, and its successes and challenges, remains largely absent from the literature (Mukherji and Shah 2005). Contrasting the lack of compelling studies of groundwater governance factors, increased groundwater pumping monitoring, and the use of geographical information systems and remote sensing have allowed for significant advances in the type of empirical methods available to researchers studying groundwater. We know a great deal generally about groundwater physical processes and interactions with human behavior, but how to translate this knowledge into specific, microlevel models of human behavioral responses is less clear (Gorelick and Zheng 2015).

For this paper, we define governance to be institutions that guide behavior and management as decisive action by an organization to reduce water extraction externalities. We use an economic framework to explain the emergence, purpose, and challenges of groundwater governance across the globe. Why does groundwater governance that fully internalizes externalities appear to be limited globally? The first explanation is supply and demand. Where water is abundant or people are scarce, issues with groundwater depletion are limited. Even by the most conservative measures of sustainability, 80% of the world's aquifers have recharge in excess of extraction (Gleeson et al. 2012). However, even where demand for water is large and supply is limited, groundwater governance is incomplete. This is the focus of our work.

Groundwater users often disagree on the extent of the problem and the necessary solutions (Ayres et al. 2018); users face considerable challenges in monitoring the unobserved resource and linking cause and effect (Bredehoeft 2011); and weak institutions limit the effectiveness of many plans to reduce drawdown (Schlager 2007). We examine ten basins, located on six continents, which vary in terms of intensity and type of water demand, hydrogeological properties, climate, and social and institutional traditions via an integrated assessment along three dimensions: characteristics of the groundwater resource; externality problems; and governance institutions. Our framework suggests that the coordination required to limit aquifer depletion is costly and that these costs limit the completeness of the solution that is ultimately implemented (Demsetz 1967; Libecap 1989).

Background

Groundwater users share a resource where each pumper can reduce the water available to neighboring wells, decreasing future availability and increasing pumping costs. Under open access, pumpers do not consider the full opportunity cost of applying water to a different purpose or at a different time (Provencher and Burt 1994). Pumping creates local areas of drawdown, pulling water from the surrounding formation (Bredehoeft et al. 1982; Brozović et al. 2010; Guilfoos et al. 2013; Edwards 2016). Overlapping pumping effects can reduce or eliminate the flow of water to pumps (Merrill and Guilfoos 2017; Peterson and Saak 2018).

Economic-hydrologic models help predict where institutional controls on extraction may be beneficial. The early groundwater economics models suggested pumping externalities were limited, governance provided limited benefits, and given the expense, was unlikely to be undertaken (Gisser and Sanchez 1980; Burness and Brill 2001). However, altering the initial assumptions increased modeled gains from management, for instance changing marginal pumping costs (Gisser and Sanchez 1980; Guilfoos et al. 2013); allowing exhaustion of groundwater (Merrill and Guilfoos 2017); adding well capacity constraints (Foster et al. 2015; Manning and Suter 2016; Foster et al. 2017); or the addition of ecologically valuable outflows (Esteban and Albiac 2011).

Additional insight has been obtained by incorporating lateral subsurface groundwater flows using Darcy's Law and the Theis Equation into economic models (Saak and Peterson 2007; Brozović et al. 2010; Suter et al. 2012; Guilfoos et al. 2013). Lin Lawell (2016) notes the recent shift in economic-hydrologic modeling to incorporate spatial work that addresses the complex relationship between groundwater users and the movement of groundwater. Koundouri et al. (2017) also highlight the importance of spatial analysis to groundwater economics, which incorporates heterogeneity in aquifers from well placement, crop choices, and physical properties such as recharge, saturated thickness, or hydraulic conductivity.

However, work translating modeled gains from coordinated pumping into an understanding of groundwater governance has been limited. In cases where resource heterogeneity creates varying local conditions and preferences, “national governmental agencies are frequently unsuccessful in their efforts to design effective and uniform sets of rules to regulate important common-pool resources across a broad domain” (Ostrom 1999). Subsurface groundwater systems are complex, and achieving first-best management outcomes is difficult. Yet localized policies that are more efficient and garner more popular support may provide politically feasible second-best solutions (Edwards 2016; Guilfoos et al. 2016).

The primary issue in groundwater depletion is human behavior, not geophysical processes. Groundwater provides two challenging issues to incentivize cooperative behavior in conserving water. First, the impacts of pumping and external costs are generally unknown or difficult to ascertain. Second, the majority of benefits in conserving a depleted aquifer accrue in the future, and therefore the incentives are stacked against cooperative action in the present. Without institutions to help with economic incentives, human behavior in an open-access resource will often be myopic (Ostrom et al. 1992; Walker and Gardner 1992; Ostrom et al. 1994). While it is also possible that groundwater pumpers behave strategically (Pfeiffer and Lin 2012; Liu et al. 2014; Oehninger and Lawell 2019), more evidence is needed to establish the extent of spatially strategic behavior. A game-theoretic model of pumping behavior does not seem to comport with behavior in a lab, where information is more easily available and stochastic processes can be controlled (Suter et al. 2012).

We focus our analysis on broadly defined aquifer systems across six continents, as shown in figure 1. For the remainder of the paper we combine the adjacent but hydrogeologically separated Indus and Ganga-Brahmaputra basins due to their similarity in geography, hydrologic properties, and governance issues, leaving us with ten basins for analysis. We chose these systems to provide a diverse cross-section of locations, hydrogeological properties, and governance regimes, while including some of the world's largest systems in the most water-scarce areas, and aquifers important to irrigated agricultural production. Section 3 provides an integrative framework for comparing the basins. Section 4 describes the basins in detail and section 5 draws general conclusions based on our economic framework. In section 6 we discuss the challenges to effective groundwater governance and how future research linking economic and hydrologic understanding could be important in solving these problems. Section 7 concludes.

Details are in the caption following the image
Groundwater basins used in this study and worldwide irrigation intensity. Source: Author's drawing using data from the Food and Agriculture Organization of the United Nations, UNESCO, the European Environment Agency, and ESRI [Color figure can be viewed at wileyonlinelibrary.com]

Integrative Framework

In this section we develop a framework for systematically evaluating groundwater governance. Figure 2 shows a two-cell groundwater model that can incorporate the basic dynamics found in economic-hydrologic models. The left panel depicts the physical setting, which is symbolically represented in the right panel. The figure can be imagined to expand to more cells to create a connected grid, each cell with a different user. In a model composed of parcels i = 1,2,.,n overlying the groundwater basin, each user has a net benefit function from groundwater that includes the revenue earned and the costs of pumping. The net benefit function for user i across the projected life of the aquifer (T) is:
NB i w i h i = 0 T π i w i h i e δt dt ()
Details are in the caption following the image
Two-cell groundwater model with two users [Color figure can be viewed at wileyonlinelibrary.com]
Net benefits are an increasing function of groundwater pumped, w, and the elevation of the groundwater, h. We initially assume that groundwater has low salinity and groundwater use does not change the water quality of the remaining stock directly. We suppress the time subscripts for simplicity. More water pumped results in greater revenues, although marginal revenue may be declining. A higher water table elevation decreases the pumping lift and thus lowers cost. Since groundwater extracted from the aquifer by all users affects the future state of the groundwater resource, we describe the change in elevation in the state equation:
h ˙ i = r i w i θ h i h i ()

Recharge, r, groundwater extracted, w, and transfers of water between parcels determine the change in elevation on a given parcel. The flow of groundwater is determined by the elevation differential of groundwater levels at neighboring parcels, hi − hi, where the notation -i indicates all surrounding parcels. The volume of transfer is determined by the properties of the soil and saturated thickness of the aquifer at the parcel, which is captured by the coefficient θ.

Groundwater Resource

The extent and geophysical parameters of groundwater basins vary widely across the world. To get a sense of the differences between basins we look at the key geophysical and economic variables from Equations (1) and (2).

Recharge

The supply of water available for extraction in each groundwater basin is defined by the existing stock of water and the annual influx of recharge, r. More precipitation is generally linked to greater recharge, but the ratio of precipitation to recharge varies by basin and season (Jasechko et al. 2014). Basins that accumulate negligible amounts of natural recharge are considered non-renewable and extraction resembles mining. Basins that receive nonnegligible amounts of natural recharge are renewable resources, but groundwater mining can occur where extraction exceeds recharge. Within basins, recharge can occur in one area and then flow to another and different portions of an aquifer can be considered renewable and nonrenewable.

Lateral Flows

How groundwater moves through the ground is defined by the hydraulic conductivity of an aquifer, which, along with cross-sectional area, determines the coefficient θ. Conductivity depends on soil/rock type. The major types of aquifers are (i) sand and gravel (ii) sandstone (iii) karst (iv) volcanic and (v) basement (Margat and Van der Gun 2013). The large groundwater systems we investigate here often include multiple types of sediment. Sand and gravel aquifer systems characterize the most commonly exploited systems and have high porosity and permeability, which lead to greater flow velocities for groundwater (Margat and Van der Gun 2013).

Groundwater Elevation

The elevation of groundwater is determined dynamically by the existing stock of water, current recharge into and flows out of the aquifer, and anthropogenic extractions. Groundwater systems are either confined or unconfined. Confined aquifers have impermeable layers that keep the aquifer pressurized. Unconfined aquifers have a water table that is at atmospheric pressure and is more likely to be connected to surface water. Groundwater systems can maintain surface flows during dry times of the year and affect water supply downstream. Heavy extraction of groundwater may reduce baseflow in streams when ground and surface water are connected (Mukherjee et al. 2018). Groundwater elevation is seasonal in many aquifers. In basins that are connected to river systems, surface water flows during dry seasons reduce groundwater elevation. The seasonal pumping of water itself also creates cones of depression around high-producing wells that can interfere with neighboring wells, with the drop in the water table limiting or eliminating production at neighboring wells during irrigation season.

Extraction

Demand for groundwater is characterized by the type of economic activity the water is put to, and along with the cost of pumping, determines the amount of extraction, w. Water demand is typically broken into irrigation, manufacturing and industrial, and public supply (water for drinking and sanitation). In many of the aquifers explored in this paper, depletion occurs most severely due to irrigation demand. However, depletion is a state of the physical system determined by the properties of the groundwater system as well as water use decisions, and is not equivalent to water demand. Changes in the efficiency of water use, for instance improving irrigation practices, can change water demand functions without necessarily reducing physical depletion (Grafton et al. 2018).

Externalities

We modify the model in Equation (1) to include an externality term, E(·) that is a decreasing function of the elevation of the groundwater table:
NB i w i h i = 0 T π i w i h i E i h i e δt dt ()

Equation (3) explicitly separates externality costs from an individual user's pumping costs, although both costs are decreasing in h. The equation of motion remains Equation (2). Under open access conditions, a myopic pumper will ignore the effect of pumping on resource stock, causing aquifer drawdown and increasing the pumping cost in all future time periods for nearby wells; as well as decreasing the availability of water to buffer fluctuations in precipitation. These externalities are referred to as the pumping cost and the risk externalities, respectively, and are captured in the original specification in Equation (1). The reduction in water table elevation also has the potential to cause additional externalities, of which we focus on six: loss of artesian pressure; depletion of surface flows; land subsidence; local areas of rapid drawdown; bedrock interactions; and seawater intrusion. Externalities are ordered arbitrarily since the importance of the externalities is local in nature and differs by circumstance. Figure 3 provides a detailed explanation of these six externalities.

Details are in the caption following the image
Diagram of groundwater pumping externalities. Notes: The left panel shows the natural groundwater system without pumping. While most groundwater systems are likely to be in flux within a season and over time as conditions change, a static equilibrium is assumed for illustrative purposes. A confined aquifer provides subsurface pressure for a free-flowing artesian well, while the stock of fresh groundwater prevents the saltwater aquifer from moving inland. When pumping begins, the reduction in the elevation of the confined aquifer lowers the elevation of the piezometric surface, reducing or eliminating the artesian pressure in the well (W1). Pumping creates a cone of depression (W2) which lowers the water table in the immediate vicinity of a stream, reducing surface flows. When several wells are located in close proximity, (W3, W4, W5) they create overlapping cones of depression and areas of rapid depletion. Land subsidence occurs when high density pumping results in aquifer drawdown, causing the potentially permanent compaction of aquifer rock and sediments, lowering surface elevation, and damaging infrastructure. Additionally, rapid pumping can cause water tables near a well to decline to the point where they interact with bedrock, or otherwise become too deep for economical extraction. Pumping near the coast (W6) can break the natural barrier freshwater aquifers provide, leading to costly brackish water intrusions into coastal wells

Governance

As in many common-pool settings, establishing effective governance institutions to solve the problems of open access are costly. Complete property rights to groundwater are seldom defined because of high measurement and bounding costs (Burness and Brill 2001). Whether a basin engages in collective management is determined by the size of expected gains; number of bargaining parties; heterogeneity of bargaining parties; information on the resource and its value; and the concentration or skewness of the allocation system (Libecap 1993). Groundwater basins can be conducive to collective management, however, because they generally have well-defined boundaries with a limited and relatively stable group of overlying landowners who may already engage in other repeated cooperative actions (Ostrom 1990).

Our framework for understanding groundwater governance is one of net benefits and transactions costs (Ayres et al. 2018). More stringent controls are observed when the value of the resource is high or with lower costs of bounding, monitoring, and enforcement of rules on extraction or use (Demsetz 1967). Transaction costs, which are the costs of defining, enforcing, and trading property rights (Coase 1960; Allen 2011), can render beneficial controls too expensive on net. To formalize this process, assume users can undertake a policy that protects the water table by conserving some amount of water for each portion of the aquifer, relative to open-access drawdown. Each user bears a cost of this policy of κi. The new conservation pumping path is ωi, where ωi ≤ wi, and as a result users see higher groundwater elevations, ηi ≥ hi. Each user sees net benefits from the new regime if:
N B i ω i η i κ i N B i w i h i ()
In aggregate, the benefits of the policy are positive when:
i NB i ω i η i κ i i NB i w i h i ()

As external costs increase, so do the aggregate benefits of conservation pumping regime. Conservation policies where the inequality in Equation (5) is satisfied do not necessarily satisfy the inequalities in (4) for all users, or users' perceptions of (4). Therefore, the distribution (and perceived distribution) of the benefits and costs of potential management approaches determines how difficult implementation will be. Satisfying the inequality of Equation (5) is not sufficient to guarantee collective action as other aspects of social choice can be important, like fairness.

Formal modeling of this process is difficult because of the level of specificity required to understand policy alternatives and their translation to individual perceptions of benefits and costs. To model this complexity would require examining transaction costs as a function of specific cultural, legal, and political factors, for instance setting the total transaction costs as something like ∑iκi(Z), where κi maps a set of specific governance factors, Z, into a numeric cost for each user. Empirically, the factors underlying total transaction costs are largely unobservable or difficult to quantify at the aquifer scale. We abstract away from these complexities by asserting that governance involves comparing management strategies on both benefits and transaction costs.

The spatial nature of groundwater externalities determines local benefits and costs and is crucial in understanding the observed institutional response. Spatial externalities mean problems are typically localized; local solutions reduce the number (and often heterogeneity) of users who need to coordinate on a solution. However, spatial externalities can increase transaction costs where they create heterogeneous effects. Spatial effects can also create complexities that hinder effective management.

Drawing on the tradition of transaction cost economics, we systematically link the “principal dimensions” of groundwater challenges to the “principal attributes” of management solutions and then demonstrate how the “predicted alignments are corroborated by data” (Williamson 2000). We develop a typology, shown in table 1, which categorizes management measures on their ability to control drawdown. Because the externality cost is increasing in drawdown, grouping management approaches provides an ordering across externality types. The proposed typology is intended to classify management control and graduated property rights over pumping decisions, but not the associated costs of doing so, and thus does not provide a ranking of net governance effectiveness. The tiers are typically increasing in transaction costs due to the greater controls and infrastructure needed to quantify and track water use and rights at higher tiers.

Table 1. Tiers of Groundwater Governance
Progression Description Evaluative Literature
Tier I
Open access Few or no limitation on pumping by overlying landowners/users Kanazawa (1992)
Tier II
Management entity formation Formation of districts, councils, etc. to promote conservation, define scope of problem, advocate for policy Edwards (2016); Ayres et al. (2018); Nachbaur (2007)
Well permits/entitlements Control of right to drill and maintenance of well database Guilfoos et al. (2016)
Well spacing Minimum distance requirements for new wells Edwards (2016)
Tier III
Area closure rules End issuance of permits for specific regions Edwards (2016); Ifft et al. (2018)
Well monitoring requirements Mandatory metering of wells Babbitt et al. (2018)
Well retirement Removal of wells from production Tsvetanov and Earnhardt (Forthcoming)
Groundwater recharge Investment for artificial replenishment Harou and Lund (2008)
Local uniform rules Cutbacks or pricing implemented uniformly Smith et al. (2017); Drysdale and Hendricks (2018); Huang et al. (2013)
Tier IV
Binding pumping caps Limits on total basin extraction and assignment of individual pumping caps Ayres et al. (2018); Ayres et al. (2019)
Punitive rule enforcement Monetary or other penalty for excessive withdrawals

Halder (2019);

CA Water Code 100732

Tier V
Groundwater banking Storage and ownership of recharged groundwater Guilfoos et al. 2016
Groundwater markets Transfer of pumping rights Kuwayama and Brozović (2013); Nicholas and Young (2014); Edwards et al. (2018); Wheeler et al. (2016); Manjunatha et al. (2011)

The first grouping, Tier I, combines a variety of limited management regimes that resemble open access conditions. Overlying landowners are allowed to pump, generally without binding limits. Tier II includes measures that require documentation and permitting of pumping, but generally do not affect existing users or provide a binding cap on individual extractions. This tier includes the formation of an entity to study the problem of groundwater drawdown; requirements for permits or entitlements to drill a well, typically providing a corresponding database of well locations; and spacing restrictions placed on well permits.

Tier III approaches are characterized by limiting the ability of new entrants to drill wells; issuing uniform cutbacks on pumping; requiring individual users to monitor their pumping; allowing certain wells to be retired from use; restricting or closing the issuance of drilling permits in sensitive areas or those considered overdrawn; and implementing recharge into the groundwater basin. Tier IV includes more significant individual regulation of pumpers, including binding caps on pumping and rule enforcement. Measures in this tier provide regulatory authority to reduce pumping but do not provide individual monetary incentives to conserve water or replenish the aquifer. The Tier V grouping includes the most complete tracking of water via assigned and tradable property rights, allowing groundwater markets and banking.

Because our modeled externalities are all caused by groundwater drawdown, solutions will typically involve limiting pumping. However, areas of rapid drawdown and bedrock interaction are also tied to well proximity and may be more easily resolved via well-spacing rules and areas closures. Externalities with a less localized spatial component, like land subsidence and loss of artesian pressure, respond more readily to monitoring and uniform cutbacks. Externality type can also impact transaction costs, with issues related to seawater intrusion and reductions in surface flows often requiring heterogeneous action among a subgroup of users, which can increase transaction costs. We return to the question of governance tiers and externality types in the empirical discussion in section 5.

Background on Study Basins

We apply our integrated framework to groundwater governance issues in ten basins, which we describe in more detail below. Table 2 provides a summary of the basins, including the primary countries overlying the basins, basin area, and the maximum thickness of the water-bearing formation. Where available, we have also included measures of the cumulative volume of depletion of each aquifer, the time period for which it is measured, and groundwater footprint. The groundwater footprint is “the area required to sustain groundwater use and groundwater-dependent ecosystem services,” with measures greater than one characterized as unsustainably exploited (Gleeson et al. 2012).

Table 2. Long-Term Groundwater Depletion in Major Aquifers
Basin Primary Countries Extent (x 1,000 km2) Max Thickness (m) Reference Period Cumul. Depletion (km3) GF/AA
Renewable
Ogallala United States 450 150 1900–2008 353 9.0
Central Valley, CA United States 80 600 1900–2008 113 6.4
North China Plain China 320 1,000 1900–2008 170 7.9
Northern India Systems India, Pakistan, Nepal, Bangladesh ~920 600 1900–2008 1,361 18.4–54.2
Guarani Argentina, Brazil, Uruguay, Paraguay 1,200 800 * * <1
Mancha Aquifers Spain 12.76 Thru 2002 3.1 **
Calama Chile 0.6–0.8 210 * * **
Non-Renewable
Arabian Aquifer System Saudi Arabia >1,485 6,500 1900–2008 468 38.5–48.3
Nubian Aquifer System Egypt, Libya, Sudan, Chad 2,176 3,500 1900–2008 98.4 **
Great Artesian Basin Australia 1,700 3,000 1880–1973 25 <1
  • Notes: GF/AA is a measure of groundwater stress based on the groundwater footprint over the area of extraction (Gleeson et al. 2012). Estimates are based on mean runs of data and ** are aquifers not classified, from Gleeson et al. (2012). *Cumulative depletion numbers are not available for Guarani and Calama Aquifers. Sources: Margat and Van der Gun (2013), Konikow (2011), Foster and Loucks (2006, p.82), Esteban and Albiac (2011), and Jordan et al. (2015).

Indus and Ganges-Brahmaputra Aquifers

The Indus Aquifer in northwestern India and the Ganges-Brahmaputra Aquifer in northeastern India and northern Bangladesh are heavily exploited, primarily for agricultural irrigation. In India, there are few regulations on groundwater use and extraction costs are heavily subsidized. Some areas, like Rajasthan, have passed legislation that sets up a state groundwater authority, restricts tubewell development, and provides a framework for privatization, but corruption and a lack of trust by local farmers of state-run institutions hinder success (Birkenholtz 2009). Bangladesh, like India, suffers from subsidies and cheap extraction costs, as well as high levels of groundwater development with weak and underdeveloped management institutions (Qureshi et al. 2014). Groundwater depletion has caused reduced surface flows in the Ganges-Brahmaputra basin, disrupting irrigation.

North China Plain Aquifer

The North China Plain Aquifer is located in a major floodplain and is extracted heavily for irrigation water. The deep parts of the aquifer are confined and essentially receive no recharge, creating concerns about groundwater mining. Recently the use of double cropping in some areas has led to drops in groundwater levels of 1 m per year (Yang et al. 2017). Depletion has led to concerns over seawater intrusion near the coast and land subsidence further inland.

Calama Basin Aquifer System

The Calama aquifer is located within the Loa River water system in northern Chile's Antofagasta Desert and supplies a majority of the region's municipal, agricultural, and industrial water supply, including water to the region's mining industry, which supplies 17% of the world's copper supply (Cochilco 2014; Edwards et al. 2018). Due to high demand in mining, the opportunity cost of freshwater in the region is extremely high, with mining firms paying to desalinate and pump water to their mines when groundwater is unavailable.

Guarani Aquifer System

The Guarani is one of the largest transboundary aquifer systems in the world, underlying parts of Argentina, Uruguay, Paraguay, and Brazil. The Guarani system has an immense stock of water, is not yet heavily exploited, and receives considerable recharge, making governance less pressing than for other systems. An international agreement on aquifer management was signed in 2010 between the four countries, but is limited in its scope and powers (Villar and Ribeiro 2011).

Arabian Aquifer System

The Arabian Aquifer System is a sandstone aquifer primarily underlying Saudi Arabia. Due to an arid climate, it is nonrenewable and has experienced heavy depletion due to large-scale production of wheat in Saudi Arabia from the 1980s to the late 2000s. This has led to a large loss in groundwater reserves (Elhadj 2004), although there has been a shift in water policies in recent years (Ouda 2014). The loss of groundwater reserves is the primary externality, though pumping costs (Müller et al. 2017), water quality (Rehman and Cheema 2016), and seawater intrusion are also important externalities.

Nubian Aquifer System

The Nubian Aquifer System (NAS) is a sandstone aquifer in the Northeast of Africa underlying the Sudan, Egypt, and Libya. It is the largest fossil water aquifer in the world and is primarily used for irrigated agriculture. Total groundwater extracted in this aquifer system has increased greatly from the 1970s (Konikow 2011). In some areas of the aquifer, extraction has led to salination concerns and, due to the low amount of recharge, concerns about future groundwater availability for irrigation and drinking water. Oases and sabkhas can be dried up by the extraction, as natural outflows from the aquifer dwindle.

Great Artesian Aquifer

The Great Artesian Basin, the largest artesian basin in the world, has faced increasing problems as bore wells were drilled to access the pressurized water within sandstone deposits. The growth of borehole wells has resulted in the diminishing of natural discharge with some natural springs drying up entirely (Habermehl 2006). In 1996, the Groundwater National Reforms required all boreholes to be licensed, and in 2000, the Australian government adopted a strategic plan for the basin that has led to improved governance and restrictions on water extraction from wells, set back requirements, and trading programs. It is noteworthy, however, that licensed wells are not necessarily monitored wells, and many issues with measuring and limiting the use of groundwater persist (Wheeler et al. 2020).

High Plains Aquifer

The High Plains Aquifer in the central United States underlies portions of eight states and provides extensive irrigation water for crop production in Texas, Oklahoma, Kansas, and Nebraska. Governance institutions are at the state or local level and include well-spacing and area closure rules to address areas of rapid depletion (Edwards 2016); market exchanges to address stream depletion (Kuwayama and Brozović 2013); and uniform pumping reductions to address long-run overdraft (Drysdale and Hendricks 2018). Pumping externalities, local exhaustion of stock, and depletion of surface flows are the primary externalities across the basin.

Central Valley Aquifer

The Central Valley Aquifer underlies California's Central Valley, one of the most productive agricultural regions in the world. The basin is somewhat unique relative to the others in this study in the large amount of surface water flowing into the basin. Heavy exploitation of the aquifer as both a primary and supplementary water source for crop production, including high-value perennials, has created extensive drawdown and land subsidence. Some areas of the aquifer have shown subsidence of up to ten inches per year (Jeanne et al. 2019).

Mancha Aquifers

The Mancha aquifers are two geographically adjacent aquifers in geologically distinct river basins. Variable precipitation cycles since the 1970s have led aquifer extractions to draw down water tables in dry years (Apperl et al. 2015; Closas et al. 2017). The primary externalities associated with aquifer drawdown in Spain are the depletion of water for future supply, the impairment of surface flows, and reduced pressure in artesian wells.

Cross-Country Synthesis

Table 3 provides an economic comparison of the ten aquifer systems in our study. For each aquifer we compute the ratio of extraction to recharge as a measure of the extent to which use is sustainable. The Arabian Aquifer has minor amounts of recharge, but is generally classified as nonrenewable. Of the renewable aquifers, all except the Guarani see depletion in excess of recharge, meaning their water tables have been drawn down over time. We do not have estimates of discharge in each of these basins, which prevents the calculation of a complete water balance. However, we do compute a rough use-to-stock ratio to provide information on the duration of time for which water reserves would be available, at current levels of extraction.

Table 3. Economic Factors and Observed Governance of Select Aquifers
Basin Recharge (net km3/yr) Use (km3/yr) Stock (km3) Use/Recharge Ratio Use/Stock Ratio Externalities Governance
Ogallala, 6–8 ~17 15,000 2.43 0.12% LD, SF, BR II-V
Central Valley, CA, 7 ~11 1,130 1.57 0.97% LD, SU II-III
North China Plain,, 49.2 6,000 LD, SU, SI II
Northern India,, 176 230 31,000 1.31 0.74% LD, BR, RI I-II
Guarani, 45–55 1.0 30,000 .02 0.003% NA I
Mancha Aquifers 0.6 1.01 10.5 1.68 9.6% SF,AP II-III
Calama 0.2 0.3 1.5 SF V
Arabian- 1–2.76 16 2,185 5.8 0.62% LD,SI II
Nubian, −0.2 2.2 14,470 NR 0.015% RI,SF I-II
Great Artesian, ~1 0.55 8,700 NR 0.0063% AP I-III
  • Notes: The table provides referenced estimates for recharge, use and stock that may be highly uncertain and contain significant rounding areas. The Use/Recharge Ratio (use divided by recharge) and Use/Stock Ratios (use divided by total stock) are calculated by the authors based on the numbers in the table. Recharge numbers are not always clearly characterized as net of depletion, and estimates in this table may or may not subtract out depletion. Externality and governance data are based on author assessments of existing literature on each aquifer and externalities are selected to represent primary externalities of concern, although other externalities may be present.
  • * Authors were unable to find a clear estimate of use for the North China Plain Aquifer.
  • ** An estimate of the stock of water in the Calama Aquifer is not currently available.
  • Externalities: AP, loss of artesian pressure; BR, bedrock interactions; LD, local areas of rapid drawdown; NA, not currently problematic; RI, loss of buffer to mitigate risk; SF, depletion of surface flows; SI, seawater intrusion; SU, subsidence.
  • 1 Sources: Konikow 2011.
  • 2 Foster and Loucks (2006).
  • 3 Margat and Van der Gun 2013.
  • 4 Calama Basin numbers are highly uncertain and come from government reports DGA (2003) and DGA (2005) as reported by Edwards et al. (2018) and Jordan et al. (2015).
  • 5 Sindico et al. (2018).
  • 6 Foster et al. (2009).
  • 7 Apperl et al. (2015) and Esteban and Albiac (2011).
  • 8 Richey et al. (2015).
  • 9 Herczeg (2008).

Generally, it is the renewable systems with large use-to-recharge ratios that see the highest tiers of governance, pointing to the governance level increasing in potential benefits of solving externality problems. India is a key exception. India faces a difficult governance challenge due to relatively weak underlying institutions and high transactions costs as a result of large numbers of users. In addition, heterogeneity at the local level as well as geographically broader differences in management options for shallow and deep aquifers suggest key transaction costs limiting coordinated governance action (Fishman et al. 2011; Blakeslee et al. 2020).

Each aquifer in the table, apart from the Mancha Aquifers, has a use-to-stock ratio under 1%. This implies that on average, at current pumping rates, full depletion would take more than 100 years. Management is generally not designed to prevent water from “running out.” Instead, local externalities described in the seventh column of table 3 drive the level of governance characterized in the eighth column. Surface water externalities appear the most likely to be noticed and addressed. The highest tiers of governance correspond to areas where groundwater pumping has disrupted surface flows: the Ogallala, Mancha, and Calama aquifers. Well-defined surface water property rights often legally require groundwater pumpers to take some action to reduce aquifer depletion. Conversely, the Central Valley aquifer, which also has a high recharge ratio, has seen limited governance of its key externalities related to drawdown and subsidence. Like the North China Plain and areas of the Ogallala, local drawdown externalities have been addressed using Tier II-III management approaches. And, despite the potentially high cost of land subsidence, basins facing this issue have not adopted high-tier levels of governance as a result of a misalignment of benefits and costs. In contrast to the depletion of surface flows, land subsidence does not generally lead to a firm legal claim against extractors, and thus pumpers, even as a group, do not bear the full costs of land subsidence.

Nonrenewable systems also appear to have lower tiers of governance, which can be attributed to their limited interaction with surface water. The Great Artesian Aquifer has relatively more stringent management, and it is notable that despite being nonrenewable, water is naturally discharged to the surface throughout the basin. The focus of management has been on water use efficiency and capping unused boreholes to limit the loss of natural springs (Habermehl 2006).

Transaction costs can limit the transition to higher tiers of management, and these costs appear to differ with the type of externality. When surface water is depleted by groundwater pumping, surface property right holders have legal recourse, lowering uncertainty and transaction costs. Where transaction costs are high, users have used management approaches that economize on transaction costs when broad actions are infeasible (Williamson 1981). For example, agreements to cap pumping are often prohibitively expensive, but the importation of water to supplement pumping or recharge the aquifer is viewed favorably by pumpers. Examples include groundwater recharge in the Central Valley, importation of water to the North China Plain, and the desalination and pumping of water in Calama.

Governance is a process of balancing costs and benefits, and good governance does not necessarily require the progression from less stringent to more stringent management. For instance, it does not appear the Guarani Aquifer needs tradeable permits and monitoring to address externalities. In addition, arriving at a high tier does not resolve many management issues or necessarily indicate effective governance. On the Calama Aquifer, Chile's 1981 water code, which established tradable water rights that were separable from the land, is supposed to limit extraction to the level of recharge, allocate extraction rights to individuals, and facilitate full water trading. However, the original water code included few environmental controls (Bauer 1997). As a result, mining water extraction has reduced water tables and surface flows, leading to declines in arable land and wetlands. Recent policy changes have statutorily enforced restrictions on transfers that could further lower water tables and the basin is generally closed to new pumping rights (Hearne and Donoso 2014; Edwards et al. 2018).

Based on the intracountry heterogeneity in groundwater management approaches, we conclude that externalities and transaction costs, rather than national or regional government type or national or multinational water policy, are the key determinant of groundwater governance choices. Within Spain, the US, Australia, and Chile the success of groundwater governance varies across and even within groundwater basins. Centralized, authoritarian governments do not appear to have had success in limiting groundwater externalities. In China, recent changes to enact a permit system with the potential for water right transfers (Zheng et al. 2012) have been hindered by poorly defined rights and the poor performance of local institutions (Chen et al. 2018). In Saudi Arabia, state planning has been the cause of, rather than the solution to, groundwater depletion.

While local, nested institutions recognized by higher levels of authority can be effective (Ostrom 1990), these institutions are weakened or not present under authoritarian regimes. In contrast, the High Plains Aquifer in the United States also saw ineffective governance and open-access drilling in the 1960s, but groundwater users in Kansas petitioned the state for local control, forming five groundwater management districts that initially implemented well spacing and area closures, but which have over time implemented local areas of uniform cutbacks to address surface water depletion and local areas of sustained drawdown (Edwards 2016; Drysdale and Hendricks 2018). While emerging research from China suggests there is potential for locally initiated water organizations (Chai and Zeng 2018), the issue of decentralized control under authoritarian regimes remains a key hurdle (Yu et al. 2016).

Discussion

In this paper we provide a framework in which the benefits and costs of collective action guide the choice of groundwater governance options. The tiers of groundwater governance are chosen endogenously, with both the benefits and costs of collective action affecting observed governance. The goal of governance is to address externalities or reach specific policy goals of resource users. Moving up the tiers may or may not be beneficial. In fact, for the same benefits, the lowest cost governance rules would be the best choice.

There are potential shortcomings of comparing large groundwater basins across the world. Information is imperfect and many of the groundwater estimates are made with significant uncertainty. Data is collected with different methods based on the country of origin. By focusing on large units for analysis we may gloss over important local governance institutions within subbasins. For instance, the Mancha aquifers described in this paper have surprisingly different governance outcomes, despite sharing many similarities (Esteban and Albiac 2011).

Another challenge in comparing groundwater governance institutions is that many rules and approaches can be used concurrently. The goal of groundwater governance is not to reach the highest tier in our framework, or to be solely self-governing; top-down and bottom-up approaches can be complements. Threats of top-down restrictions may result in more efficient bottom-up efforts to enact collective action. In contrast, top-down goals of extraction limitations may be challenged by local users in regions where the immediate benefits of irrigation are salient. Effectively managing long-term extraction is not the priority of a subsistence farmer with limited opportunities for earning income and growing food. Sometimes, and especially where groundwater stocks are large, sustainable governance could be a second-order concern; in northern China, heavily extracted aquifers have reduced poverty (Huang et al. 2006).

Cultural differences and preferences also require consideration in performing comparative analysis. Rules, and the process under which those rules are made, depend on local customs. Social scientists armed with economic recommendations are likely to need domain specific knowledge of cultures and customs for successful transmission of governance knowledge. Governance also faces added challenges when it spans multiple countries, for instance the Nubian and Guarani aquifers, due to the increased transaction costs to developing institutions. Neighboring countries may differ in legal traditions and the current state of intracountry governance. However, because externalities are local, transboundary governance provides a framework for the shared resource only in the immediate vicinity of the border. It does not solve internal governance challenges near the border, and a lack of transboundary governance should not be viewed as a barrier to governance away from the border.

Empirical groundwater economics is dependent on data and modeling of the externalities to understand the potential gains from management. Models of groundwater extraction may contain detailed groundwater dynamics, but be limited in describing aspects of water demand or institutions. These models may have specific domains in which they are useful, and researchers must carefully select models when attempting to inform policy. Many subbasins suffer from multiple externalities and joint problems relating to water quality and water quantity. It is often easiest to focus modeling on one externality at a time; a key challenge is understanding when externalities and governance should be considered multidimensional. Agent-based and computational modeling holds promise to help understand the dynamics of complex systems with multidimensional issues (Gailliard et al. 2014).

More data-intensive approaches should incorporate detailed microlevel datasets into the public domain to compare local governance efforts. This requires that governments collect the relevant data. However, detailed data on well numbers and locations, as well as rates of pumping, are typically not available. AquaSTAT, from the FAO, is an example of a national-level data repository. However, more repositories that house microlevel basin information are needed. Common datasets with both physical, historical, and economic data can establish microlevel evidence for the success of different institutions under different conditions. Satellite data offers an especially promising data source for future research, as it offers the potential for consistent measurement of groundwater depletion across countries.

Another fruitful area of research is the incorporation of behavioral economics into groundwater governance research. Aspects such as framing (Menegaki et al. 2009), cognitive processes (Brozyna et al. 2018), forecasting (Tong and Feiler 2017), learning (Rodela et al. 2012), and salience to risk (McCoy and Zhao 2018) are all areas of behavioral economics that can be incorporated into groundwater research to a greater extent. Such an approach could lead to more effective strategies to describe behavior and design policies to meet the increasing challenges to groundwater governance.

Conclusion

This article provides an economic framework for understanding and evaluating groundwater governance across the globe. Economic analysis suggests aquifer drawdown is likely too rapid in many areas because of the common pool problem: Pumpers fail to internalize all the social costs of their individual pumping. Governance addresses the externalities associated with decreasing water table elevations, and in most cases these problems are defined by local conditions, i.e., both human behavior and hydrogeological properties. Our analysis of governance tiers across ten major groundwater basins shows that the value of groundwater and type of externality tend to determine the level and type of groundwater governance in a basin. Aquifer drawdown, extraction in excess of recharge, occurs in most of the basins we analyze and aquifers with larger use-to-recharge ratios tend to have higher tiers of governance. However, externalities related to surface water depletion appear to be addressed in the study basins more often, and more completely, than other externalities.

Transaction costs can limit the transition to higher tiers of management, and many of these costs decrease as more information about the physical aquifer system becomes available. Emerging hydrologic models, geospatial and remote-sensed data, and microlevel pumping and economic data offer potential tools for identifying externalities and reducing transaction costs. Critically, the analysis of groundwater management approaches must incorporate the transaction costs of coordination and implementation. Governance cannot be evaluated strictly based on what type of management program has been implemented. A tier-V regime is not necessarily better than a tier-III regime because there are costs of implementation. Instead, effective governance balances management approaches that reduce common pool losses with the transaction costs of creating and running the regime. Analysis of governance effectiveness requires local hydrological, institutional, and social context.

Acknowledgments and/or disclaimers

The authors would like to thank participants at the American Geophysical Union Fall Meeting, and attendees at the Seminar in Water Economics Online (SWELL) for feedback on earlier versions of this paper. We also thank Terry Jordan, Manuel Pulido Velázquez, and Jose Albiac for sharing information on physical and governance characteristics of specific aquifers and David Sutherland for help in creating figures. This work was supported by the USDA National Institute of Food and Agriculture, Hatch project 1,017,720 and multistate project 1,020,662.

    • 1 The model's results rely on the assumption of a bathtub-like basin, accessed by many pumpers, and where pumping at one location is instantaneously transmitted to others.
    • 2 Other externalities, water quality issues and pollution from overland activities, are excluded primarily for brevity and simplicity.
    • 3 Ostrom (1990) suggests that the governance process itself determines in part the benefits and costs. For instance, the involvement of stakeholders in developing institutions is critical to promote prosocial behavior and to get buy-in from the users that they are designed to govern.
    • 4 For instance, seawater intrusion near Monterey, CA created divergence in the incentives to strictly regulate benefits (pumpers near ocean) and costs of reducing pumping (all users). Inland farms not affected by seawater intrusion refused to join coastal users in limiting pumping to protect coastal drinking water supplies (Ayres et al. 2018).
    • 5 For example, to address surface depletion externalities, limiting groundwater pumping near the stream or river is effective. In Nebraska, groundwater trading to account for these dynamics required a more complex management approach to account for differences in the transmittance of pumping effects, but this spatial heterogeneity also increased the benefits of more secure institutions and transferable water rights, relative to uniform basin-wide management approaches (Kuwayama and Brozović 2013; Brozović and Young 2014).

    The full text of this article hosted at iucr.org is unavailable due to technical difficulties.