Contested baselines and transboundary water resources management, with illustrations from the Nile

ABSTRACT

The effect of a policy intervention is estimated as the difference between a state of the world without the intervention – the dynamic baseline – and with the intervention. Negotiations on managing transboundary water resources confront three kinds of problems in specifying the dynamic baseline: unexamined baselines, uncertain baselines and contested baselines. This paper focuses on contested baselines. Controversy may arise from different ethical or political assessments of the appropriate choice of the state of the world without the policy intervention. The Grand Ethiopian Renaissance Dam on the Nile highlights the importance of understanding contested baselines involving transboundary water resources.

KEYWORDS:

• Transboundary water management
• baselines
• Grand Ethiopian Renaissance Dam
• water policy
• Nile basin

Acknowledgements

The author thanks Mohammed Basheer, Thinus Basson, David Fuente, Giorgio Guariso, Jim W. Hall, Kingsley Haynes, Chris Heady, David Hulme, Marc Jeuland, Martin Keulertz, Jack Knetsch, Lisa Robinson, Aaron Salzberg, V. Kerry Smith, Eelco van Beek, John Waterbury and Kevin Wheeler for helpful comments on drafts of this paper.

Disclosure statement

No potential conflict of interest was reported by the author.

Notes

1. For this work he was awarded the Stockholm Water Prize in 2008.

2. If the policy alternative only has very short-term effects, then a dynamic baseline may not be required; status quo conditions may be sufficiently accurate as a comparator. However, the management of transboundary water resources typically requires a long planning horizon and a dynamic baseline is necessary.

3. For example, in a cost–benefit analysis the discount rate is often a parameter that affects the estimate of the net benefits of shifting from the dynamic baseline to a new state of the world with the policy intervention.

4. For ex-post evaluations, various non-experimental methods are available for the construction of the dynamic baseline, such as synthetic controls, even if there is only one (or a few) treated units (Abadie, 2021).

5. For simplicity, Figure 1 shows the lives lost after the construction of the dam to be constant. This need not be the case. Technological innovation, population and economic growth, and climate change could cause lives lost to rise or fall after time t.

6. This might occur because the dam encouraged settlement below the dam and increased the number of people at risk from the release of excess flood waters or dam failure.

7. Sometimes a cost-effectiveness measure is reported as the inverse of this ratio, that is, the dollars required for a unit change in the performance indicator, $/ΔPI. In this case, the dynamic baseline is hidden in the denominator of the ratio.

8. New approaches may be required for thinking about decision-making under deep uncertainty (Lempert et al., 2013).

9. Careful consideration of uncertainty in the forecast of the dynamic baseline also focuses attention on what is known or being assumed about the magnitude of the treatment effect. Is the magnitude of the treatment effect estimated as (1) a percentage change in the value of the performance indicator in the dynamic baseline, or (2) an absolute magnitude change over a specified baseline? If the treatment effect is estimated as a percentage change in the dynamic baseline, then the size of the change will be dependent on the specific forecast of the dynamic baseline. If it is measured as an absolute magnitude, then it may be possible to apply this estimate to different forecasts of the dynamic baseline. Which approach to use is an empirical question and context specific. For most RCTs, the treatment effect is typically reported as a percentage point change in the performance indicator between the treatment and controls, not as percentage change in the dynamic baseline. Whether this percentage point change can be applied to another baseline is a question of the external validity of the experimental results.

10. I thus make a distinction between contested baselines and Kinchy’s notion of contentious baselines. Contested baselines refers to disputes over future dynamic baselines; Kinchy’s notion of contentious baselines refers to disputes over past or status quo conditions.

11. Conceptually, endogenous policy decisions could be incorporated into the dynamic baseline in such an analytical model, but policy analysts typically avoid this approach because it is so challenging. However, stakeholders may not be able to avoid confronting this complexity.

12. Conceptually this problem of contested baselines is similar to the use of hypothetical baselines in stated preference studies in which respondents are provided with a description of a baseline that is intentionally not the actual state of environmental quality, health or other baseline condition. The stated preference researcher then poses a valuation question or choice task that is contingent not on the existing status quo, but rather on the state of the world described in this new hypothetical baseline (Whittington & Adamowicz, 2011; Whittington et al., 2017). Just as with contested baselines, there may be disagreements among respondents about the appropriate specification of the hypothetical baseline in a stated preference survey due to perceived inequalities in the status quo conditions.

13. Knetsch (2020) refers to both the willingness-to-accept measure to value a loss and the willingness-to-accept measure to value a gain that is viewed as a reduction in a loss as ‘the domain of losses’ (see also Nguyen et al., 2021).

14. There was also disagreement between Egypt and Sudan over the appropriate baseline. Egypt argued that the baseline should be based on current water use: Sudan felt that the dynamic baseline should be based on the water allocations in the 1959 Nile Waters Agreement because it was not yet using its full water allocation under the treaty.

15. Sudan has two main concerns about the construction and operation of the GERD. First, the elimination of the Nile flood would eliminate the remaining recession agriculture along the Nile in Sudan. Second, Sudan needs timely information on the operation policy of the GERD in order to better operate its downstream dams, especially Roseires.

16. Deltares does water resources modelling throughout the Nile basin, indeed throughout the world, and has had a long relationship working with Egypt on Nile water management issues.

Additional information

Funding

This work was supported by the UK Research and Innovation Economic and Social Research Council [ES/P011373/1] as part of the Global Challenges Research Fund through the “Future Design and Assessment of water-energy-food-environment Mega Systems” (FutureDAMS) research project.