Techno-Economic Assessment of Non-Revenue Water: A Case Study at AER, GWCL

This study utilizes multiple regression analysis and the technique for compu-ting economies of scale to evaluate the dynamics of NRW in the AER of GWCL. Data on monthly total production, billed consumption, total revenue, total production cost and the volume of NRW spanning the period January 2015 to June 2019, was obtained from the headquarters of AER and used for the analysis. The study showed that NRW averaged 40.3% for the period under consideration, which far exceeds the 25% threshold set by the World Bank for developing economies. It was also established that a fairly inelastic relationship exists between NRW and total revenue. Results further show that resources are not optimally used in the AER as proof of diseconomies of scale was observed.


Introduction
According to a Water Rights report [1], freshwater constitutes about 2.76 percent of the total water available on Earth. Global freshwater demand is projected to increase by 55% by the year 2050 [2]. This makes freshwater a very scarce resource, which needs to be well managed by man.
A large number of countries are experiencing water stress, and increasing drought and desertification is already worsening these trends. Approximately one-third of countries have medium or high levels of water stress. Almost all countries that have registered high water stress are located in North Africa and West Asia or in Central and South Asia, and these levels indicate serious water difficulties in the supply of freshwater, at least during parts of the year [3]. By 2050, it is projected that at least one in four people will suffer recurring water shortages. One of the targets of the Sustainable Development Goal six (SDG 6) is to substantially increase water-use efficiency across all sectors and ensure sustainable withdrawals and supply of freshwater to address water scarcity and substantially reduce the number of people suffering from water scarcity by 2030. To ensure universal safe and affordable drinking water, there is the need to invest in adequate infrastructure, provide sanitation facilities, and encourage good hygiene practices.
In Ghana, portable water is supplied by the Ghana Water Company Limited (GWCL), formally known as the Ghana Water and Sewerage Cooperation (GWSC) [4]. The company mainly provides water to the urban centers of the country and currently covers about 77 percent of the urban areas in Ghana. It supplies 192 million gallons of water per day, through eighty-eight (88) urban water supply systems, which is in excess demand of 57 million [5]. To ensure a sustainable production of water to its customers and to meet its projected targets, the company operates fifteen (15) regional offices which are distributed around the country. The offices include, the Accra East, Accra West, Ashanti South, Ashanti North, Ashanti Production, Western, Eastern, Brong Ahafo, Northern, Volta, Central, Upper East, Upper West, ATMA Production and Tema.
The GWCL assumed its current structure after it underwent a reformation period in the late 1990s. Prior to independence, the need to have a division to supply water to the colonial apparatus operating in the country at the time was One of the major reasons for privatizing the GWSC was the issue of Non-Revenue Water (NRW) and its impacts on the operations of the water supplying organization [4] [6] [7]. Conceptually, NRW is water that has gone through all the production process but lost without generating revenue to the utility company.
According to the International Water Association [8], NRW is "the difference between System Input Volume and Billed Authorized Consumption". Greater Accra region (see Figure 1 for location of Accra East Region). The dynamics of NRW have not been established for the Accra East Region which constitutes the largest GWCL administrative region in terms of urban water demand. Thus, this study, therefore, seeks to fill the gap in terms of information on NRW dynamics by conducting a technical and economic assessment of Non-Revenue Water (NRW) to ascertain the trends and its effect on production and

Theoretical Framework
The magnitude of Non-Revenue Water (NRW) has grave consequences on the competitiveness and efficiency of water production and distribution companies.
Large volumes of NRW constrain the financial ability of utility companies to embark on strategic projects such as network expansion, meter replacement and metering policies as well as improving operational efficiency. Furthermore, attempts to minimize the magnitude of NRW to acceptable levels require significant financial investment in fixing leakages and improving the database of customers as well as strengthening monitoring systems [7] [11].
Moreover, reducing NRW has significantly positive impact on the financial performance of utility companies. According to World Bank report [11], an economic optimum is achieved when the marginal cost of reducing non-revenue water exceeds the marginal benefits (water savings). Benefits derived from reducing NRW should be measured through decreased cost of production [11].
This is because reduction in NRW could lead to reduced volumes of water production to meet the demands and supply requirements of the service area. Besides, additional revenue could be generated through the sale of additional water captured from reduction in the volume of NRW [12] [13] [14] [15].
A study conducted on the evaluation of apparent water losses [16] emphasized that the reduction in the cost of production through reduction in NRW has little financial incentives for water service providers if volumetric tariffs are low (flat rate) and few or no metering systems available. Thus, from the economic perspective, minimizing NRW leads to decreased cost of production, increased volume of billed water and increased revenue generation. However, efforts must be made by utility service providers to charge economic rates on water usage as well as improve on the metering system in service areas [12] [15]. Thus, the implicit theoretical function for analysing NRW is shown in Equation (

Trends in Non-Revenue Water
The first objective of the study is to analyze the trends in Non-Revenue Water in the Accra East Region of the GWCL. The growth rates of non-revenue water are estimated to enable the utility company to realize its non-revenue water over time. In a discrete form, growth rate is calculated as the difference between current values and previous values all divided by the previous value. The result is C. K. K. Sekyere et al. Journal of Water Resource and Protection then multiplied by 100% to provide an interpretation in percentage terms. The algebraic framework generating the growth rate is espoused as: where: η = the growth rate of Non-Revenue Water; NRW t = Non-Revenue Water at time t; NRW t−1 = Non-Revenue Water at time t -1.

Elasticity of Non-Revenue Water to Total Revenue and Production
Elasticity measures the degree of sensitivity or responsiveness of a variable to another variable under consideration [16] [17]. The coefficient of elasticities provides a statistic to determine whether a variable is elastic or inelastic. To examine the elasticities of NRW to total revenue and production, two separate log-linear regression models are specified [17] [18]. These are expressed as follows: The elasticities are calculated as follows: where: TR γ = the elasticity of Non-Revenue Water to total revenue; Q γ = the elasticity of Non-Revenue Water to total production; ln t NRW = the natural logarithm of Non-Revenue Water at time t; ln t TR = the natural logarithm of total revenue at time t; ln t Q = the natural logarithm of total production at time t; t ε and t µ are the stochastic error terms; i β and i α are the parameter estimates.
A variable is said to be elastic if its coefficient of elasticity is greater than one.
However, it is said to be inelastic if the coefficient of the variable is less than one [16] [17]. For instance, when a proportionate change in NRW leads to the same proportionate change in total revenue, the relationship is considered to be unitary elastic.

Economies and Diseconomies of Scale
Economies of scale are a long-run phenomenon in the production process of a firm. A firm enjoys economies of scale if expansion in production results in lower unit cost of production. Also, diseconomies of scale exist when expansion  [19]. Advanced Microeconomic Theory [19] opined that economies of scale are derived by finding the inverse of the elasticity of total cost to total production.
To derive the estimates of economies of scale, a quadratic cost function is specified in Equation (2.7). The quadratic cost function is preferred over the linear cost function because the former captures the concept of diminishing returns in the production process [17] [19]. Stated differently, the linear cost function assumes that cost increases at a constant rate per unit increase in production whiles the quadratic function assumes non-constant rate of increase in cost with production increases.

Data Type and Source
Monthly time series data was used for the study. The data was collected from the headquarters of the Accra East Region of the Ghana Water Company Limited (GWCL). The data runs from January 2015 to June 2019. The data employed includes monthly total production (m 3 ), billed consumption, total revenue, total production cost and volume of non-revenue water (m 3 ).

Summary Statistics
Stata 14 was used to analyse the data. Table 1  revenue performance of the Accra East region of GWCL. Indeed, the revenue collection ratio has been estimated at 76% [20]. This shows that several private consumers of water as well as government agencies owe GWCL.

Trend of Non-Revenue Water in the Region
This section analyses the trend of NRW between the 1 st month of 2015 and the 6 th month of 2019. NRW is computed as the difference between total volume of treated water and the volume of billed water expressed as a percentage. The results obtained from this analysis reveal an undulating trend of NRW over the study period. It is estimated that NRW percent averaged 40.3% over the period under consideration (see Figure 2). It can also be seen from Figure 2 that the highest NRW% recorded was in December, 2016, where 53.03% of total production    [20]. This indicates that African countries have a lot to do to reduce NRW in their economies.

Trend in Water Production and Sales
The study further examines the trend of volume of water produced and the volume sold to customers (see Figure 3). The results obtained show a fairly increasing trend for the production of water in the Accra East Region of GWCL.
As expected, due to population growth, urbanization and increased use of water, GWCL has expanded the productive capacity to meet increasing demand. It is important to note that 2,589,310 m 3 of water was produced in January 2015.
However, production has increased substantially to 4,819,183 m 3 in April 2019, which represents an increase of 86.1% over the period. It is important to note that the growth of water production significantly exceeded growth in water sales to customers. Essentially, the difference between production and sale of water is referred to as the volume of NRW as indicated by the bars in Figure 3. It can be observed from Figure 3 that the gap between water production and sales (NRW) steadily widens over the period under consideration. This shows clearly that GWCL has to put in measures to minimize the gap.

Trend in Production Cost, Billing and Revenue Collection
This subsection examines the trend of cost of production, billing of water sold and the revenue collected. The results obtained (see Figure 4) show that an average of GH¢1,984,720 (equivalent to $367,336.66) has been expended by GWCL for the production of water between January 2015 and June 2019. It can be observed that the cost of production had fairly increased over time (see Figure 4).
This trend can be attributed to increased volumes of water production and inflationary tendencies.  Table 2 presents the elasticity coefficient of NRW to total revenue (output of Equation (2.4)). The model is statistically significant at all levels with a coefficient of determination (R 2 ) of 83%. This result has two implications. Firstly, the model can be used for policy analysis, and forecasting total revenue. Secondly,

Elasticity of Non-Revenue Water to Total Revenue and Production
the results indicate that about 83% of the variations in total revenue can be explained by variations in NRW and production of treated water [21] [22]. The estimates show that one percent increase in the volume of NRW decreases total revenue by 0.032%. In other words, total revenue and NRW are negatively related. This finding is statistically significant at 5%. It is important to note that values of NRW were captured as non-positive (see Table 2) as they do not yield revenue for GWCL. However, the estimates indicate that the relationship between the two variables is fairly inelastic. Specifically, a proportionate increase in NRW reduces total revenue by less than a proportionate amount [4] [16] [17].
Also, the study reveals that production is positively related to total revenue. A percentage increase in the volume of production of treated water increases total revenue by 1.78%. This estimate is statistically significant at all levels. Likewise, this estimate is elastic. This is because increase in production will cause total revenue to increase if volumes produced are not captured under NRW. However, more revenue will be generated if GWCL sells increased volumes of treated water to customers at the prevailing tariffs [16] [23].
Furthermore, the study estimates the elasticity of NRW to total production.
The results captured in Table 3 are principally the output generated from equa- The results obtained indicate that one percent increase in NRW increases total production of treated water by 0.32% [24]. In other words, total production and NRW are positively related. This result is statistically significant at the 1% level.
The results further indicate that the relationship between total production and NRW is fairly inelastic. Indeed, [25] emphasized that reducing NRW could lead to decreased volumes of water production to meet the demands of the service area. With the assumption that households have relatively stable demand for

Type of Economies of Scale Present in GWCL
The elasticity of production to total cost of production is presented in Table 4.
The estimates indicate that a percentage increase in production increases total cost by 0.93%. This shows that a fairly inelastic relationship exists between production and cost. This finding is statistically significant at 1%. However, the squared term of production is statistically insignificant. Based on the estimates, Essentially, a proportionate increase in all inputs in the production process translates into a less than proportionate increase in production of treated water.
In other words, resources at the disposal of the Accra East region of GWCL are not optimally employed.

Conclusions
Non-Revenue Water has been a bane in the quest to engender efficiency, sustainability of water production and management and competitiveness in the operations of the Accra East region of GWCL. The most significant deduction that can be made from the study is that NRW averaged 40.3% over the period under consideration, which far exceeds the 25% threshold accepted by the World Bank for developing economies at the Accra East region of the GWCL. This result is Based on the evidence of shortfalls suggested by the findings of this research and conclusions arrived at, the study makes the following recommendations: • Monitoring and evaluation practices should be intensified by GWCL to minimize the volume of NRW in the region. Task force from GWCL should periodically conduct checks on households to ascertain the accuracy of meter readings, issues regarding meter bypass and illegal connections in communities. To this effect, GWCL should develop and assemble watchdog committees in various communities to help clamp down on illegal connections and issues of meter bypasses. This is because, these individuals live in the communities and have evidence of households who engage in these malpractices. • The Government of Ghana should also design policies that would clear debt owed the GWCL and pay for its share of unbilled authorized consumption.
An agreement should be reached between GWCL and government which would ensure meter installations at government agencies which do not pay for water. This would engender responsible water use and reduce the volumes of NRW.
• The GWCL should also embark on extensive meter installation project to replace dilapidated meters and also abolish the system of flat rate billing. These initiatives would minimize inaccuracies in meter registration and ensure that households that use more water, pay more. Effective implementation of this strategy would improve on revenue generation of GWCL.
• Finally, efforts must be directed at retrieving monies owed GWCL by customers. It has been evident from the study that revenue collection has been appalling within the Accra East region of GWCL. Efforts must be made to institute a prepaid system of water usage to minimize debts owed GWCL. A pilot study can be adopted in one of the districts to ascertain the performance of the prepaid system. Upon successful implementation, the exercise can be