2.3.2. Baseline Model Inputs

We use annual data from the World Development Indicator (WDI) database published by the World Bank [21]. The data spans a 50-year period, from 1970 through 2019. In setting up the variables in the system, we follow [22]. The full set of variables in the system are ordered as follows:

$$(variables)\equiv \begin{pmatrix} \textit{algorithm} \; \textit{land} \\ \textit{food} \\ \textit{female} \; \textit{agric.comployment} \\ \textit{male} \; \textit{agric.comployment} \\ \textit{food} \; \textit{agent.comployment} \\ \textit{food} \; \textit{export} \\ \textit{consumption} \; \textit{per capita} \\ \textit{investment} \; \textit{per capita} \\ \textit{real GDP} \; \textit{per capita} \\ \textit{CPI} \; \textit{inflation} \\ \textit{unemployment} \; \textit{rate} \end{pmatrix}$$

(8)

The variable "agricultural land", which is measured in squared kilometers, is the share of land area that is arable and used for permanent crops and permanent pastures. The variable "food production" refers to the production index of food crops that are considered edible and nutritious. The variables "food import" and "food export" are, respectively, the share of import and export of food items (as defined by Standard International Trade Classification (SITC) sections 0, 1, 4, and SITC division 22) in total merchandise trade. The variable "female agric. employment" represents the share of the female labor force employed in the agricultural sector. Likewise, "male agric. employment" is the share of the male labor force engaged in the agricultural sector. When appropriate, we substitute these two employment variables for a single variable termed "agric. sector employment", which represents the share of Egypt's total labor force that is employed in the agricultural sector. Similarly, we substitute food production for "agricultural sector output", which sums up the value added of agriculture, forestry, and fisheries. We note that the term "agricultural sector output" and "agricultural production" are used interchangeably in this paper. Moreover, consumption per capita, investment per capita, real GDP per capita, CPI inflation, and unemployment rate are as defined in standard literature. Although the data generally spans from 1970 to 2019, we do acknowledge missing entries in some of the series, especially for the earlier part of the period. We therefore make adjustments as we see appropriate (Appendix A).

#### 2.3.3. The Transmission of GERD in the Macroeconomy

The loss in Egypt's available agricultural land due to the construction of GERD is expected to cause significant disruption in the supply of food and other agricultural products, which will have significant ramifications for the broader economy. To examine the extent of this disruption, we first carry out sensitivity analysis where we examine the responses of selected economic variables to changes in available agricultural land. We then use the estimated elasticity coefficients and the projected losses in available agricultural land (Table 1) to map out the potential losses/gains in the economic variables under the alternative filling scenarios.

First, we estimate the elasticity coefficients for agricultural sector output and food production in response to changes in the size of agricultural land. The quantitative frameworks for estimating these elasticities are given as follows, respectively:

$$
\ln - A \text{g} \text{ri} \mathbf{Y}\_t = \alpha\_Y + \beta\_Y \ln - A \text{g} \text{ri} \text{land} \, d\_t + \gamma\_Y X\_{Y,t} + \varepsilon\_t \tag{9}
$$

$$
\ln\text{In\\_Food\\_Prod}\_t = \alpha\_F + \beta\_F \ln\text{...}\,\text{Agriland}\_t + \gamma\_F X\_{F,t} + \varepsilon\_t \tag{10}
$$

where the variables *ln\_AgriYt* and *ln\_Food*−*Prodt* represent Egypt's agricultural sector output and food production in logs, respectively, whereas *ln\_Agrilandt* represents the size of Egypt's agricultural land, also in logs. *β<sup>Y</sup>* and *β<sup>F</sup>* are then the elasticity coefficients between agricultural sector output and food production on the one hand, and Egypt's available agricultural land on the other hand, respectively. *XY,t* and *XF,t* are vectors of other independent variables (in logs) which include Egypt's population, private investment, and private consumption. Given the composition of variables in Equations (9) and (10), we conduct model fitness tests through a Variance Inflation Factor (VIF) analysis. The VIFs and the associated tolerance levels (not shown) suggest the presence of multicollinearity. To correct for this, we drop private investment and consumption from (9) and (10). This confines *XY,t* and *XF,t* to the size of Egypt's population.

Moreover, a Breusch–Pagan (BP) test [23] reveals that the baseline model is not robust to heteroscedasticity (with a reported chi-square statistic and *p*-value equal to 13.66 and 0.0002, respectively). We therefore employ robust cluster procedures in order to obtain estimates that are consistent even in the face of the heteroscedasticity. Estimating Equations (9) and (10) yields *β<sup>Y</sup>* = 0.33 and *β<sup>F</sup>* = 0.73, which are both statistically significant at the 1% level. Having obtained the elasticity coefficients for agricultural sector output and food production, we then regress other macroeconomic variables on agricultural sector output in order to obtain the elasticity coefficients for these variables with respect to agricultural production.

#### **3. Results and Discussion**

#### *3.1. Losses in Egypt's Water and Agricultural Land*

Figure 3 and Table 1 show the relationship between average annual losses in Egypt's Nile water allocation and GERD filling scenarios. Given the GERD reservoir volume of 74 km3, we estimate losses in Egypt's annual water allocation to be 51.29 ± 2.62%, 24.75 ± 2.76%, and 18.78 ± 2.76% for the 3, 7, and 10-year filling scenarios, respectively. The nature and extent of the loss depends on the length of time it takes for Ethiopia to fill the reservoir. Losses that emanate from shorter filling horizons are expected to be more severe on impact, but less persistent, whereas losses from longer filling horizons are expected to be relatively less severe on impact, but drag on for longer periods.

Egypt will lose 52.75 ± 2.44%, 28.14 ± 2.56% and 22.61 ± 2.58% of their agricultural land, relative to the baseline, for the 3, 7 and 10-year filling scenarios, respectively (Figure 4; Table 1). The losses in agricultural land presented in Figure 4 and Table 1 are calibrated under the assumption that there is no mitigating strategy in place by the government of Egypt. These estimates therefore represent the worst-case scenario in terms of losses in agricultural land. In subsequent studies, we plan to extend the analysis to include various mitigation strategies that are likely to be implemented by the government of Egypt.

#### *3.2. Projected Trends in Selected Economic Variables: Baseline Model*

Table 2 presents a numerical summary of projected trends in selected variables in a baseline (no GERD) scenario over a 3-year period. In this scenario, Egypt's agricultural sector output is projected to grow at an average annual rate of 2.27 ± 0.71% for the next 3 years, with minimum growth expected to be 1.47% and maximum to be 2.83%. Moreover, consumption per capita, which is a measure of overall welfare to some extent, is projected to grow at an annual rate of 4.74 ± 2.12% in the 3-year horizon.

**Table 2.** Projected average annual growth (%) in selected variables in a baseline (no GERD) scenario over the next 3 years.


In addition, real GDP per capita is projected to grow at an average annual rate of 2.38 ± 1.65%. This translates into an average annual growth in real GDP of 4.51 ± 1.65%, with minimum growth in the 3-year horizon projected to be 3.43% and a maximum 6.41%. Real GDP growth rate in Egypt pre-pandemic was estimated to be 5.6%. This declined to 3.6% during the fiscal year 2019/2020, following the COVID-19 shocks. However, the World Bank projects that, assuming vaccines are steadily rolled out through 2021 and early 2022, Egypt will start regaining its pre-pandemic growth momentum by the fiscal years 2021/2022/2023 [8]. Similarly, the IMF projects Egypt's GDP growth to ramp up from 2.5% in 2021 to 5.8% by the year 2025 [24]. This puts our estimates of real GDP growth right within the range projected by both the World Bank and the IMF. In what follows, we examine deviations from these trends arising from GERD under the alternative feeling scenarios.
