Does political uncertainty affect water resources development? The case of the Eastern Nile

Abstract

In water resources, there is a long tradition of utilization of methods to address hydrological and economic uncertainty. Less frequently considered, however, is how uncertainty rooted in political factors such as power asymmetry, the strength of institutions, and the interests of stakeholders, contributes to decision-making. This paper explores political uncertainty and its interaction with more routinely considered forms of uncertainty in international river basins. Using the example of the controversies surrounding major new infrastructure projects in the Eastern Nile Basin, we show that political uncertainty may play a key role in shaping the decisions of individual riparian countries about how to proceed with water resources development. Specifically, we consider whether uncertainty over the prospect of future cooperation might help explain why seemingly optimal economic outcomes that require cooperation (due to interconnectedness) remain elusive. We conclude with reflections on other dimensions of the water resources planning problem – climate change, changes in regional development matters, and preferences – that similarly require a framework that accounts for political uncertainty.

Keywords:

• Nile Basin
• Uncertainty
• Hydroeconomic optimization models
• River basin governance

1 Introduction

The stochastic nature of meteorological processes and the unpredictable dynamics of future demographic, economic and political conditions create multiple layers of uncertainty for riparian countries sharing international river basins. Surface water resources are highly variable in space and time; hydrological uncertainty in international river basins therefore creates particular management challenges (Loucks, Van Beek, Stedinger, Dijkman, & Villars, 2005; Ostrom, Burger, Field, Norgaard, & Policansky, 1999). In addition, because water resources are essential for so many human activities (Hanemann, 2006), the value of the resource to riparians depends critically on a broad range of factors that span varied economic sectors, such as population growth, technological innovations, water use efficiency, economic growth, market fluctuations, energy demand and supply, and economic development strategies. These issues may spill across national or other jurisdictional boundaries (Rogers, 1991), and may link with activities that take place outside of river basins (Berck, Robinson, & Goldman, 1991).

In the field of water resources planning and management, there is a long tradition of utilization of systems models that account for hydrological and economic uncertainty, and for the complex interlinkages between water and economic systems. Water resources engineers, for example, have developed sophisticated methods (such as various dynamic optimization models) to handle hydrological uncertainty (Sahinidis, 2004; Stedinger, Sule, & Loucks, 1984). Sensitivity analysis, decision-tree analysis, and Monte Carlo simulation are among the tools often used in dealing with economic uncertainty.

In comparison, the effects of political uncertainty – another major source of uncertainty in international river basins – are rarely explicitly considered (for exceptions, see Lowi, 1995; Mirumachi, 2015; Pahl-Wostl, 2002). The sources of such uncertainty arise from major events in the political arena, such as elections, military coups, and collapse of government (Gleditsch, Furlong, Hegre, Lacina, & Owen, 2006), as well as more gradual shifts in domestic priorities and discourse or international relations (Feitelson, 2002; LeMarquand, 1976). Political uncertainty may have significant implications for the functioning of resource management institutions (Giordano, Giordano, & Wolf, 2005), and may therefore play a key role in determining the prospect of cooperation among basin riparians. For example, new political leaders emerging from an election may denounce water treaties negotiated by their predecessors, and governments of upstream riparian countries may decide to operate their river-regulating facilities (such as dams and reservoirs) to gain strategic advantage over downstream riparian countries, once such facilities are in place.

Several characteristics of political uncertainty make it particularly difficult to predict and to mitigate, in comparison with hydrological and economic uncertainty. First, the patterns of political uncertainty involve myriad factors that are difficult to characterize and that may not overlap well with river basin boundaries (Islam & Susskind, 2012; Warner, Wester, & Bolding, 2008). Political processes, both in international relations and domestic affairs, can be extremely volatile and occur in non-recurring patterns, or may be highly path dependent and irreversible (Pierson, 2000). Second, basic behavioral assumptions about governance may be challenged in the presence of political uncertainty. Governments may behave irrationally in certain circumstances, engaging in activities that are harmful not only for other riparian countries but also for their own citizens (Shepsle & Bonchek, 2010). Third, unlike the case of hydrological and economic uncertainty, there is no robust mechanism to understand and mitigate the effects of political uncertainty on water resources development because the nature of such impacts is determined by government action, as are any uncertainty-mitigating mechanisms (Koremenos, 2005). To put it simply, governments cannot insure against the risks associated with their own actions.

Prospect theory may help to understand the impact of political uncertainty on decisions by individual riparian countries in water resources development projects (Kahneman and Tversky, 1979). Due to the difficulties in predicting political uncertainty, the fear of losses may outweigh potential gains for riparian countries contemplating new infrastructure projects. Such loss aversion may generate a strong status quo bias. Prospect theory may also help to explain why decision-makers would gravitate toward strategies that enable them to minimize potential losses in reputation and credibility (Levy, 1992). Egypt's long-standing position in opposing dam construction in upstream catchments of the Nile Basin may be due to fear of potential losses given the difficulties in assessing and addressing water retention and abstraction once such dams are built.

The omission of political uncertainty by analysts of international river basin dynamics may lead to unrealistic expectations about prospects for cooperation among riparians. In fact, the very mechanisms designed to mitigate hydrological and economic uncertainty may increase vulnerability to political uncertainty. For example, the consideration of hydrological uncertainty alone may suggest that there are compelling incentives for riparian countries to cooperate with each other to share benefits and overcome the adverse impacts of uncertainty. Enacting solutions such as large-scale river-regulating facilities or water treaties and river basin organizations might appear especially helpful for reducing such uncertainties (Fischhendler, 2004; Leary, 1999). In reality, however, joint development of large-scale river-regulating facilities has been rare, and the long-term sustainability of water treaties or other institutions cannot be taken for granted due to one or more parties’ concerns about their implementation (Bernauer, 2002; Schmeier, 2014).

The reality of political uncertainty, and its complex interplay with the hydrological and economic uncertainties that play out over the long term, may result in so-called ambiguity or deep uncertainty (Groves & Lempert, 2007; Jeuland, 2009). Such political uncertainty could thus have major implications for conflict and cooperation in international river basins. Here we use the term “ambiguity” to refer to a situation where riparian countries are uncertain of each other's future intentions and actions, and about the future results of those actions. In particular, we consider that ambiguity arising from political uncertainty, as well as its interplay with the unknown evolution of future economic and hydrological changes, may influence individual riparian countries’ approaches toward cooperation.

To explore this issue more concretely, this paper considers a range of scenarios designed to explore three dimensions of uncertainty – hydrological, economic, and political – and the extent to which they may affect cooperation among the three riparian countries in the Eastern Nile (Egypt, Sudan, and Ethiopia). The Nile Basin is an interesting case due to longstanding efforts to establish a cooperative agreement among its riparian countries (Salman, 2013), even as some of these countries have pursued unilateral and uncoordinated dam development with infrastructures such as Merowe Dam (in Sudan) and the Grand Ethiopian Renaissance Dam (GERD) (in Ethiopia) (McDonald, Bosshard, & Brewer, 2009; Whittington, Waterbury, & Jeuland, 2014). Specifically, we consider whether uncertainty over the politics of cooperation might help explain why seemingly optimal economic outcomes that require cooperation (due to interconnectedness) remain elusive. We show that the resulting ambiguity has important implications for the distribution of benefits from new development projects, and argue that it may therefore play a role in the current discourse over development in this basin.

We conclude with reflections on other dimensions of the water resources planning problem – changes in climate, regional development matters, and preferences of key policy-makers – that similarly require a framework that accounts for political uncertainty. A better understanding of the relationship between these varied uncertainties and the benefits from water resources development thus seems critical to understanding the long-standing and intensifying conflicts that affect many international river basins today.

2 Methodology

2.1 The Nile Basin

Measured at 6700 km, the Nile is the longest river in the world. Estimates of the annual flow of the river range between 84 and 92 billion cubic meters (bcm), depending on the period that is examined (Blackmore & Whittington, 2008). The basin is shared by 11 countries: Egypt, Sudan, South Sudan, Ethiopia, Uganda, Kenya, Tanzania, Burundi, Rwanda, Democratic Republic of Congo, and Eritrea (Fig. 1). Few other major river systems traverse so many and such diverse countries, and the basin is famous for the varied salience and interests of riparians in the Nile (Waterbury, 2002). Yet water conflicts in the Nile Basin have some characteristics in common with conflicts along other international rivers. For one, there is a big gap between the quantity of water available in the basin and the amount of water sought by individual riparian countries for water resource development projects. For example, the government of Egypt has plans to irrigate an additional 5 million acres by 2025, which would require an additional 20 bcm of water annually from the Nile. In Sudan, irrigation and hydropower projects already on the drawing board would demand a quantity of water well beyond its current annual allocation under its 1959 Agreement with Egypt (Jeuland, 2009; Knott & Hewett, 1994). And Ethiopia and upstream riparians around the Equatorial Lakes of East Africa, which hitherto have used only limited water from the Nile, have increasingly ambitious irrigation plans to meet the needs of a dramatically increasing population. If all the plans on the drawing boards were enacted, the annual water deficit in the Nile Basin would probably exceed 50 bcm.

Fig 1. The Nile Basin.

Disagreements over water in the Nile are also influenced by features unique to the Nile Basin, such as downstream countries’ high dependency on the water, the drastic disjuncture between contribution to and utilization of the Nile water among the chief riparian countries, and the historical dominance of Egypt in political and military power despite its unfavorable geographic position at the downstream end of the system. Egypt contributes essentially nothing to the flow of the Nile, but it depends upon the Nile for 97% of its water supply, and currently consumes more than 80% of all Nile water. Ethiopia, in the uplands, contributes 85% of the water flow in the Nile Basin yet uses almost none of that water for irrigation. Historically, Egypt was able to use its relative power advantage to safeguard its access to a large share of the Nile water. This advantage is increasingly being challenged, however, as upstream countries’ have developed economically and grown in population, and as political changes have affected the region (Whittington et al., 2014). Despite these sociopolitical tensions, the Nile Basin offers significant prospects for water resource conservation and development. Since the U.S. Bureau of Reclamation first conducted a study of hydropower potential in Ethiopia in 1964, it has been known that the Nile Basin has some of the most attractive hydropower sites in the world. More recent studies have confirmed and elaborated on this work (Blackmore & Whittington, 2008; Jeuland & Whittington, 2014; Guariso & Whittington, 1987). Table 1 suggests that hydropower development (much of it outside the Nile Basin) could in fact become an avenue to economic growth for several upstream riparian countries.

Table 1. Irrigation and hydropower potential of Nile Basin countries.

Country	Irrigation potential (000 ha)	Irrigation area (000 ha)	Hydropower potential (MW)	Installed capacity (MW)
Burundi	185	14	1366	36
Congo	NA	11	530,000	2829
Egypt	4434	3266	3210	2825
Eritrea	NA	28	NA	NA
Ethiopia	3637	190	162,000	378
Kenya	352	67	30,000	611
Rwanda	160	4	3000	59
Sudan	4843	1946	1900	225
Tanzania	828	190	20,000	339
Uganda	202	9	10,200	155

In addition, much Nile water is presently lost to evaporation and seepage as it flows north toward the Mediterranean, but such losses could be lowered if storage levels in the downstream system could be reduced (Blackmore & Whittington, 2008; Jeuland & Whittington, 2014). The Main Nile flows through harsh desert lands, where net evaporation and seepage losses are substantial in comparison to southern more tropical and highland reaches of the river. If more water were withdrawn by upstream riparians, storage at Lake Nasser would fall, reducing evaporation losses, and enabling increased water withdrawals in most years.¹ But such increased system-wide withdrawals in most years decrease the storage for use during droughts, and would therefore require increased cooperation during years with low flows to mitigate their effects.

2.2 Hydrological and economic modeling

In order to explore the sensitivity of the distribution of economic benefits from cooperation to various forms of uncertainty, this research applies a basin-wide hydroeconomic (HEM) model previously developed specifically for the Nile, the Nile Economic Optimization Model (NEOM) (Whittington, Wu, & Sadoff, 2005; Wu & Whittington, 2006). HEMs are needed to fully capture the complexities and inter-relationships that relate potential hydrological changes and water resource developments and management options on the one hand, to economic outcomes on the other (Harou et al., 2009). There are many different types of HEMs (Bekchanov, Sood, & Jeuland, 2015); the NEOM is one particular type that considers the production of a specific set (i.e., hydropower and irrigation) of economic benefits that are derived from the water system over an annual time scale. The NEOM user can specify different conditions – which in our application govern hydrological runoff, economic values, and water allocation rules – to compare the optimal benefits derived from these two sectors under a range of potential futures. For the work described in this paper, the model schematic was updated to incorporate recent changes in the Nile Basin (Fig. 2). As shown, the updated model contains 16 reservoir nodes (of which 7 are existing and 2 are under construction), 16 irrigation nodes, and 2 urban/industrial nodes. The two dams under construction are GERD and Rumela.

Fig 2. Schematic of the Nile Economic Optimization Model (for simplicity, only reservoir and irrigation nodes are numbered).

While the NEOM is a basin-wide model including both the Blue and White Nile, the analysis in this paper focuses on the three riparian countries in the Eastern Nile, namely, Egypt, Sudan and Ethiopia. Accordingly, we treat the flow from the While Nile as given in the existing configuration of water use patterns and infrastructure development. Such simplification allows us to focus on the relationship between these three major riparian countries, and on the most significant controversies among them.

To solve for maximum economic benefits from irrigation and hydropower, the NEOM uses a non-linear optimization algorithm that maximizes the benefits from these two sectors, which are two of the major uses of water in the Nile Basin (Whittington et al., 2005): (1)

In Eq. (1)(1) , the capitalized terms ( and ) are decision variables that represent the water quantity used for irrigation (in cubic meters, or m³) and the amount of energy production (in kilowatt-hours, or kW-h), respectively, during month t and at node i in country c. The economic value of one unit of irrigation water (in US$/m³) is represented by the parameter , and the economic value of one unit of hydropower (in US$/kW-h) is indicated by the parameter . Importantly, Eq. (1) can be maximized for the basin as a whole, or for specific countries, by specifying the composition of the set of c countries over which the summation applies.

Key model constraints then specify in more detail the physical and economic production relationships in the system. These constraints, discussed in detail in Whittington et al. (2005), include:

(a)	flow continuity and reservoir storage-elevation relationships that also allow for evaporative and seepage losses;
(b)	requirements for satisfying municipal demands prior to allocating water to irrigation;
(c)	reservoir and power generation capacity constraints;
(d)	hydropower generation equations that depend on flow and storage levels in reservoirs;
(e)	annual storage constraints that require the storage at the end of the year (t = 12) to return to the initial level of storage at t = 0, to prevent depletion of water storage;
(f)	mathematical representations of the pattern of crop demands at different irrigation sites across the year;
(g)	a 50% net reduction in flow in the White Nile as it traverses the Sudd swamps, based on the findings of more detailed hydrological studies (Jeuland, 2009; Sutcliffe & Parks, 1999).

The model uses a monthly time step and is solved using nonlinear optimization techniques in the General Algebraic Modeling System (GAMS). It determines the values of decision variables – storage levels in reservoirs, releases from dams, average hydropower head, water allocations to irrigated agriculture, and electricity generation – across all months t that maximize economic benefits.

2.3 Model scenarios

We use the updated NEOM to consider three separate dimensions of uncertainty – hydrological, economic and political (Table 2 ).

Table 2. The three dimensions of uncertainty considered in this analysis.

	Hydrological	Economic	Political
Contributing factors	*Natural climate variability (interannual, interdecadal, etc.) *Land conversion impacts on local climate & runoff generation *Long-term drivers (e.g., climate change)	*Infrastructure cost overruns *Unknown rate of regional development (demand-side uncertainty) *Supply costs (inefficiency losses, mismanagement) *Macroeconomic drivers (global economic forces, technological change)	*Infrastructure development decisions *Benefit-sharing agreements (e.g., power, food and other trade) *Lack of agreement over property rights (e.g., water rights) *Foreign policy shifts (e.g., change in alliances or willingness to take military action)
Scenarios considered in this research	Variation in runoff: 1) Low (15% below mean) 2) Long-term mean 3) High (10% above mean)	Variation in value of energy: 1) Low (US$0.04/kW-h) 2) Medium (US$0.07/kW-h) 3) High (US$0.10/kW-h)	Variation in coordination: 1) Full (Infrastructure, power trade & irrigation allocation) (COOP) 2) Infrastructure + power trade (ELINFCOOP) 3) Infrastructure only (INFCOOP) 4) None (NOCOOP)

2.3.1 Hydrological uncertainty

Hydrological uncertainty in the Nile arises from various sources, including climate variability (Conway, 2000; Eltahir, 1996), natural and anthropogenic processes of land conversion that alter both local climate and the processes governing the timing and magnitude of runoff generation (Bewket & Sterk, 2005; Hurni, Tato, & Zeleke, 2005), and large scale drivers such as global warming (Conway, 2005; Elshamy, Seierstad, & Sorteberg, 2009). Using an annual model such as the NEOM to consider the consequences of interannual variability is not appropriate, however, given that over year storage (in reservoirs and lakes) greatly buffers against its potential negative effects. For this reason, we focus on the uncertainty associated with variation on interdecadal or longer-term (climate change) scales. Specifically, we model illustrative scenarios based on low, mean, and high runoff into the system. The mean scenario uses stochastic flows that maintain the average spatiotemporal pattern of historical runoff. The low (−15% inflows) and high (+10% inflows) scenarios reflect the range of changes in runoff that are considered likely for the basin as a whole under climate change (Jeuland & Whittington, 2014).

2.3.2 Economic uncertainty

For simplicity, we constrain our examination of economic uncertainty to variation in the economic value of hydropower generation. This is a useful sensitivity analysis given substantial unknowns over the extent to which regional markets can absorb large amounts of new hydropower, particularly in the absence of power trade agreements (International non-partisan Eastern Nile Working Group, 2015; Whittington et al., 2014). Similar to our treatment of hydrological uncertainty, we consider three scenarios: low value (where hydropower is worth US$0.04/kW-h), medium value (US$0.07/kW-h), and high value (US$0.10/kW-h). The medium value is comparable to existing electricity tariffs and to the alternative cost of firm power generation in Ethiopia (Foster & Morella, 2011), where much of this new hydropower would be produced. This medium value thus assumes that these markets would be able to absorb the large new influxes of electricity associated with additional electricity generation. The high value is close to the alternative cost of electricity generation in Sudan (Jeuland & Whittington, 2014), and is also closer to the value of energy if some production is used to meet peaking demand. The low value case represents a situation where demand for the power is below the cost of alternative generation, or alternatively where costs of transmission to more distant markets, for example to Kenya (Fichtner, 2009), would decrease its net value.

2.3.3 Political uncertainty

Finally, we model four distinct scenarios of political uncertainty that vary in the degree of cooperation among riparians of the Eastern Nile (detailed assumptions governing these scenarios are presented in Table 3). The first scenario (COOP) explores high cooperation; in this case, the riparians agree to develop the two most economically-attractive remaining potential hydropower projects in the Eastern Nile (at Mabil and Beko Abo), maintain existing rights to irrigation water allocations, and coordinate on power trade. The power trade coordination is facilitated by transmission lines that already connect the two countries (though we note that capacity upgrades would be required), and that allow Sudan to obtain power more cheaply than it can produce it. These gains for Sudan are due to the US$0.03/kW-h wedge between tariffs in Ethiopia and the alternative cost of production in Sudan, described above. We assume that Ethiopia and Sudan would split these US$0.03/kW-h benefits equally.

Table 3. Differences in assumptions governing the variation across political scenarios.

Assumption	COOP	ELINFCOOP	INFCOOP	NOCOOP
New infrastructures	Mabil; Beko Abo	Same as COOP	Same as COOP	Beko Abo; Dal; Kajbar; Dagash; Shereiq
Energy benefits	As specified in Table 1; power trade yields additional gains of US$0.03/kW-h, split equally between Sudan & Ethiopia	Same as COOP	Same as COOP; but no gains from power trade	Same as INFCOOP; energy value is also US$0.03/kW-h lower in Sudan due to higher generation cost
Target (maximum) water consumption
Municipal/industrial demand (bcm/y)	Egypt: 4.3 Sudan: 1.0	Same as COOP	Same as COOP	Same as COOP
Irrigation water target (bcm/y) ^a	Ethiopia: +1 Egypt: 51.2 Sudan: 15	Ethiopia: +1 Egypt: 51.2 Sudan: 18	Same as ELINFCOOP	Same as ELINFCOOP
Evaporative losses (bcm/y) ^b	Ethiopia: 3.5 Egypt: 10–12 Sudan: 2.5	Same as COOP	Same as COOP	Ethiopia: 3.2 Egypt: 10–12 Sudan: 4.2
Total target demand (bcm/y)	Ethiopia: +1 Egypt: 55.5 Sudan: 18.5	Ethiopia: +1 Egypt: 55.5 Sudan: 23.5	Ethiopia: +1 Egypt: 55.5 Sudan: 23.5	Ethiopia: +1 Egypt: 55.5 Sudan: 23.5

a For Ethiopia, this is in addition to ∼2 bcm/y of irrigation water use in the vicinity of Lake Tana. The COOP scenario assumes that Egypt and Sudan follow the 1959 Nile Water Agreement; i.e., Egypt maintains target water demand of 55.5 bcm/y (for municipal/industrial + irrigation), and Sudan has 18.5 bcm/y (for municipal/industrial + irrigation + evaporation from new dams).

b For Sudan, this does not include evaporative losses from Gebel el Aulia, Sennar, or the original Roseires, because these dams preceded or were accepted as part of the 1959 Nile Waters Agreement.

The second and third scenarios explore somewhat reduced cooperation. In scenario 2 (ELINFCOOP), the riparians coordinate for infrastructure development and power trade only, but upstream irrigation increases beyond existing agreements. Scenario 3 (INFCOOP) then assumes the riparians only coordinate on infrastructure development, and not on power trade or water allocations. Finally, the non-cooperation scenario (NOCOOP) assumes unilateral development of hydropower infrastructure in Sudan and Ethiopia, no power trade, and increases in upstream water abstractions. In this scenario, Ethiopia only manages to finance one additional Blue Nile dam (at Beko Abo) over the relevant planning horizon, due to the high financing burden of these large new projects. We also assume that the generation of hydropower in Sudan leads to a reduction in energy benefits of US$0.03/kW-h, due to the higher relative cost of energy productions at her hydropower projects.

2.4 Other input data

Additional data used to parameterize the model were primarily obtained from Jeuland and Whittington (2014), who developed a hydro-economic simulation model that includes all of the existing and potential new dams and irrigation projects included in the updated NEOM model. These authors obtained such information from several recent feasibility studies and master plans produced by the basin riparians (BCEOM, BRGM and ISL, 1998; Norplan-Norconsult, 2006, Norplan-Norconsult & Shebelle Consulting Engineers, 2007).

3 Results

3.1 Sensitivity of economic benefits to political uncertainty

We first consider the issue of political uncertainty (Table 4). As shown, under average flow and medium demand conditions, all three countries have an incentive to cooperate, though full cooperation is only favored by Egypt. This is due to the fact that full cooperation is needed to more effectively safeguard downstream water uses. Constraints on upstream water uses keeps water abstractions in the upstream system to levels that allow Egypt to almost fully meet her irrigation water demands, and releases from Ethiopian dams are also better coordinated to meet these targets. Without constraints that prioritize downstream water use, benefits decline by about 7% in Egypt. Complete lack of cooperation then decreases Egypt's benefits an additional 3% because Sudan builds additional dams on the Main Nile to generate hydropower. These dams contribute additional evaporative losses, and thus decrease water availability in Egypt further.

Table 4. Total economic benefits (in billions of US$/y) under different scenarios of political uncertainty.

Country	Ethiopia	Sudan	Egypt	Total
Scenario
Full cooperation (COOP)	2.99	1.94	3.00	7.94
Infrastructure and power trade (ELINFCOOP)	3.17	2.03	2.79	7.99
Infrastructure cooperation only (INFCOOP)	2.87	1.73	2.79	7.38
No cooperation (NOCOOP)	2.31	1.87	2.71	6.89

Notes: All results assume mean runoff and medium electricity price. Bold indicates the best outcome for each country under each flow scenario.

For Ethiopia and Sudan, the best results are obtained for the second cooperation scenario, in which the countries agree to build the more attractive Ethiopian hydropower projects and additionally coordinate on power trade (the ELINFCOOP scenario). Power trade generates roughly US$0.3 billion per year of additional economic benefits (assumed to be split 50–50 between Ethiopia and Egypt). For Ethiopia, full cooperation with power trade generates 4% lower benefits than this scenario, cooperation on infrastructure alone (without power trade) generates 10% lower benefits, and non-cooperation is the least attractive, decreasing benefits by 27%. The large drop in benefits for Ethiopia in the NOCOOP scenarios is due to the fact that only one (rather than 2) additional dam(s) can be financed. For Sudan, benefits decline by 5% for full cooperation, by 8% for non-cooperation, and by 15% for infrastructure cooperation without power trade. The NOCOOP scenario is relatively more attractive to Sudan than to Ethiopia, because Sudan pursues other (though less economical) options for producing electricity.

3.2 Interplay between hydrological uncertainty and political uncertainty

Economic outcomes are highly sensitive to hydrological uncertainty, but the sensitivities are asymmetric with respect to the direction of change (Table 5). A 15% decrease in runoff leads to 24–26% lower benefits, while a 10% increase in flow leads to a proportionately smaller relative increase in benefits (by 4–5% overall). In addition, decreased flow has more strongly negative effects on benefits to Ethiopia (26–30% reductions) and Egypt (27–31% reductions) than to Sudan (11–18% reductions). This greater sensitivity to flow reductions in Ethiopia comes from the decrease in hydropower produced when releases through her hydropower turbines decline. The large size of the GERD in particular means that it is more difficult to maintain storage levels when flows are significantly lower. In Egypt, the reduced flows strongly reduce water availability for irrigation. As the downstream riparian in the system, these water supply shortfalls fall squarely onto Egypt (and conversely, higher runoff mainly benefits her). In practice, the riparians might agree under cooperation to share deficits, but this would not be optimal if the economic value of irrigation water is similar across locations. This is due to the well-known water efficiency penalty that stems from the water losses during transmission from the upstream system (Ray & Williams, 2002).²

Table 5. Total economic benefits (in billions of US$/y) under different scenarios of hydrological uncertainty and political uncertainty.

Country	Ethiopia		Sudan		Egypt		Total
Scenario	COOP	NOCOOP	COOP	NOCOOP	COOP	NOCOOP	COOP	NOCOOP
Low flow	$2.1	$1.7	$1.60	$1.67	$2.2	$1.9	$5.9	$5.2
Mean flow	$3.0	$2.3	$1.94	$1.87	$3.0	$2.7	$7.9	$6.8
High flow	$3.0	$2.3	$2.00	$1.92	$3.2	$3.0	$8.2	$7.1

Notes: All results assume a medium value of energy, e.g., US$0.07/kW-h. Only full cooperation or non-cooperation is shown. Bold indicates the best outcome for each country under each flow scenario.

Sudan, on the other hand, is able to meet her irrigation water demands in all three flow scenarios, such that her lost benefits result entirely from reduced dam releases and hydropower generation. Under low flow conditions, Sudan also appears to have a slight advantage in not cooperating, unlike the other riparians who both benefit from cooperation across hydrological conditions. Sudan's incentive to cooperate across the higher flow conditions is also generally low. Sudan might therefore use this position as leverage to obtain a more favorable power trade deal that would provide her greater benefits relative to Ethiopia. Given these results, one could also conclude that convincing Sudan to share water shortfalls with other riparians under reduced water availability might prove difficult, and that Sudan may be tempted to defect from agreements over specific water allocations.

3.3 Interplay between economic uncertainty and political uncertainty

Economic outcomes are also sensitive to uncertainty about the value of energy, and here the results are proportionately similar for increases and decreases in value (Table 6). A US$0.03/kW-h increase in the value of power increases total benefits by 22–25%, while a similar decrease reduces benefits by 23–26% overall. Uncertainty over the value of power has the strongest effect on benefits in Ethiopia (changing these by 38–46% across high and low value scenarios), followed by Sudan (21–30%). This higher sensitivity for Ethiopia is due to the fact that nearly all of her benefits come from hydropower, while the lower sensitivities in Egypt (ranging from 7 to 8%) are related to the fact that most of her benefits are in irrigation. Looking across cooperation and non-cooperation scenarios, it is interesting to note that Sudan has greater incentive to not cooperate as the value of energy increases. This is because the additional hydropower dams on the Main Nile in northern Sudan become more attractive when electricity is more valuable. As the value of power declines, non-cooperation becomes relatively more costly for Sudan, however, because these alternative energy projects become less economically attractive when demand is very low.

Table 6. Total economic benefits (in billions of US$/y) under different scenarios of economic uncertainty.

Country	Ethiopia		Sudan		Egypt		Total
Scenario	COOP	NOCOOP	COOP	NOCOOP	COOP	NOCOOP	COOP	NOCOOP
Low price	$1.8	$1.3	$1.53	$1.31	$2.8	$2.5	$5.6	$5.1
Medium price	$3.0	$2.3	$1.94	$1.87	$3.0	$2.7	$7.9	$6.8
High price	$4.2	$3.3	$2.36	$2.42	$3.2	$2.9	$9.8	$8.6

Notes: All results assume mean runoff. Only full cooperation or non-cooperation is shown. Bold indicates the best outcome for each country under each flow scenario.

Combining the results from examination of these various dimensions of uncertainty, we can see that there are strong economic incentives for the countries to cooperate, especially on infrastructure development and power trade, which themselves do not induce difficult economic tradeoffs. More challenging is the issue of water allocations; Sudan can clearly utilize more water for irrigation purposes, and perhaps has strong economic incentives for wanting to defect from existing limits to consumptive uses. When the value of energy is high, Sudan's incentives to cooperate decrease somewhat, because she gains larger energy benefits from developing new (albeit more costly) hydropower projects, which add to the irrigation benefits of increased irrigation water use. Meanwhile, if Nile runoff declines, irrigation benefits make up a larger fraction of the total of Sudan's benefits, which also pushes her toward less cooperation in order to increase these benefits further.

4 Discussion

Political uncertainty in water resources systems has not been well understood, and systematic modeling has been rare. This paper considered how political uncertainty – as well as its interaction with hydrological and demand uncertainty – might affect the prospects for joint infrastructure planning and management by riparians of the Eastern Nile. The Nile Basin provides an interesting case for considering the potential influence of political uncertainty in shaping the prospect of cooperation, given the difficulties and futile efforts in establishing a cooperative agreement between riparian countries, even as unilateral and uncoordinated development have proceeded.

In analyzing the distribution of economic outcomes across a range of illustrative scenarios, we showed that Egypt has perhaps the most to gain from cooperation on infrastructure and a clear distribution of water rights. Across hydrological and economic conditions, Egypt's benefits are 10–15% (or US$0.2–0.3 billion/y) higher under full cooperation compared to non-cooperation, because there is no clear way to effectively limit upstream water abstractions and unilateral infrastructure development without cooperation. Ethiopia and Sudan also benefit from cooperation, but primarily over infrastructure development and power trade. This is due to the particularly favorable economics of hydropower production in Ethiopia. Both Sudan and Ethiopia appear to have some incentive to deviate from full cooperation, but for somewhat different reasons. Ethiopia can produce more hydropower (worth about 6% of annual benefits) if releases from dams are not coordinated with downstream irrigation water requirements. Meanwhile, Sudan would benefit primarily by increasing water abstractions beyond the constraints imposed by the 1959 Nile Waters Agreement. Finally, the incentive to defect is greatest for Sudan when the value of hydropower is high, making additional dam construction in Sudan attractive, or when hydrological flows decline, since this increases the relative value of increasing irrigation abstractions.

Based on this analysis, we can conclude that uncertainty over the politics of cooperation and the possibility of defecting from the 1959 Nile Waters Agreement might help explain why the seemingly optimal economic outcomes that would be achieved under full cooperation remain elusive in this basin. The ambiguity that results from political uncertainty and the interplay between political uncertainty and other types (e.g., hydrological or economic) of uncertainty may therefore play a role in the current conflictive discourse over development of the Eastern Nile. Of course, it must also be highlighted that the countries involved in water rights agreements in the Nile (e.g., Sudan and Egypt through the 1959 Nile Waters Agreement) have no history of deviation from these agreements.

We also note that other dimensions of uncertainty in river basins may similarly give rise to ambiguity about the value of cooperation. For example, besides simply altering water availability, climate change may affect many characteristics and features of water resources systems, including the regional and temporal pattern of water demands, the value of renewable energy production, and the relative productivity of irrigated agriculture (IPCC, 2014). Changes in regional development, population migration, and in policy-makers’ preferences (e.g., heightened concern over environmental consequences) within and across countries could similarly alter the nature of benefits derived from surface water systems. Many of these behavioral aspects are poorly understood and extremely difficult to predict. They therefore seem likely to contribute to ambiguity and potentially to impede cooperation. In this sense, targeted efforts to reduce political or other uncertainties, for example by building confidence in sustained cooperative interaction, may help to reduce ambiguity and enhance the prospects for long-term coordination among riparians.

Notes

1 In addition, the White Nile flows through the Sudd wetlands, where half of its water spills and is lost to evaporation. Yet hydrological and remote sensing studies suggest that only modest amounts of that water could be saved and that disruption of the unique ecosystem could be severe (Blackmore & Whittington, 2008; Sutcliffe & Parks, 1999).

2 Of course, if the economic value of irrigation water is higher in Egypt, this could negate the efficiency effect. Our model assumes that the marginal product of water is equal across countries, but there is little reliable data on this issue.