Introduction

Although the offer of water in absolute terms is considered abundant, Mexico is facing a serious problem with relative scarcity (^{Conagua, 2012a}). This problem reflects the inability to supply water in some basins of the country, especially those where the population has been growing over the years.

The San Juan River hydrological sub-region (HSRSJ) occupies less than 3.3% of the national territory. However, approximately 6.2% of the total population of the country lives there, generating approx. 9.3% of the country’s gross domestic product (GDP). At an international level, the degree of water stress can be measured with the Water Stress Index. This index is defined as the relationship between water use and average natural water availability, where values above 40% show serious water stress. In case of HSRSJ, it shows a value equal to 77.1%, which places it among the regions in Mexico with a higher level of stress on its water resources (^{Conagua, 2016}). In the case of serious water stress, it is essential to pass policies aimed at making the different uses of water efficient in the regions and subregions. Currently, the main sectors consuming water in HSRSJ are, in order of importance, the urban public and domestic sectors, farming and industrial (^{Aguilar-Benitez, 2017}).

The manufacturing industry is particularly important to the dynamics of the economic activity in the HSRSJ, accounting for two-fifths of its GDP. It generates jobs and goods and services for consumption by the region’s population and exports outside the hydrological subregion and the country (^{Conagua, 2012b}). Water is an essential resource for the production processes used by the industries located in the basin. In 2015 alone, approximately 99 million m³ was used for production. This volume represents 7.7% of the water that industry uses in the country (^{Conagua, 2016}). For its production process, the manufacturing industry extracts approximately 56% of water from underground sources (aquifers), which are currently overexploited and are therefore declared as closed, having registered zero water availability, causing irreversible damage to the aquifers. The industry supplements its water needs from the public urban supply, which also mainly receives water from the subsoil, with the same adverse impact on the same water bodies (^{Aguilar-Benitez, 2017}). This overexploitation of aquifers, both the one generated by the manufacturing industry as well as other economic industries, has the same harmful effects on society, such as springs depletion, disappearance of lakes and wetlands, elimination of native vegetation and loss of ecosystems, reduction of well performance, increase in extraction costs, land settling and cracking, among others.

However, like most emerging or developing countries or regions, the monetary value of using this natural resource in the production process does not necessarily reflect its scarcity, as well as the catchment and treatment costs it requires. Thus, the cost of water is not properly charged, which lowers awareness of its inefficient use, waste, or pollution (^{Dupont & Renzetti, 2001}). Likewise, in general, the State has regulated the price of water in most cities in the country. Thus, the importance of the value of industrial water has been overlooked.

Water usage in the manufacturing industry adds value to the end product or service, and this added value is the value of the water that is used in the process, which reflects the effective payment by consumers for resource usage. The purpose of this article is to calculate the additional monetary value of the end product by adding an additional square meter of water to the products that are manufactured (marginal value of water) in the HSRSJ. This is done by identifying the industries that add more or less value to the resource. Likewise, knowing the economic value of water (EVW) allows for proper and efficient management of water (^{Dupont & Renzetti, 2001}). First, it offers technically-based tools to decision-makers on matters of water for making future investments, taking into account that cities and countries continue investing in water resources, which is one of the main components in the public infrastructure budget. Secondly, economic assessment is important since it enables determining whether or not the society (in this case the companies) is in favor of water supply projects. Lastly, EVW is useful to assess the design and implementation of public policy, not necessarily in terms of investment in construction but rather more efficient water management.

There is relatively little international literature about EVW calculations for developing or emerging countries with problems of scarcity, and it is mainly focused on farming and household sectors. Within this context, this article intends to fill the gap by calculating this value for the manufacturing industry in HSRSJ, a sub-region undergoing relative water stress and constant socio-economic development.

The article’s structure, in addition to the introduction, includes six sections. The next section presents a literature review, followed by a section providing information about the study site. After this section, the methodology and results obtained are presented. Finally, in section six, the conclusions and recommendations are given.

Literature review

A series of combined resources or consumables is required for the production process of goods and services that are provided for the economy, such as equipment, labor, management, capital, space, natural resources, among others contributing to the total production value (^{Ku & Yoo, 2012}). In the case of the industrial manufacturing sector, the value this sector is giving to water resource is classified as direct use value, and is assessed according to the market price to which the industry has access (^{Marcouiller & Coggins, 1999}). However, the price of the products on the market, in this case water, does not necessarily reflect their true value. In economic theory, there are three approaches for studying and calculating this production value (^{Ku & Yoo, 2012}): 1) determining water demand; 2) studying cost structure; and 3) investigating the production structure.

In the first approach, calculating water demand consists of direct statistical regressions, trying to study the relationship of the amount of the resource consumed as a dependent variable, and independent from water price, production, work, and other socio-environmental variables (^{Rees, 1969}; ^{Oh, 1973}; ^{De-Rooy, 1974}; ^{William, 1986}). However, the theory establishes that this approach generates a bias in terms of the amount of water consumed on both sides of regression, generating inconsistent and relatively questionable results (^{Hussaina, Thrikawalaa, & Barkera, 2002}).

To try to solve this obstacle, more recent research uses the same methodology, but taking different functional forms of models (^{Malla & Gopalakrishnan, 1999}; ^{Onjala, 2001}; ^{Hussaina et al., 2002}).

The second approach -cost structure- relates the total costs of the activity as a dependent variable and the prices of different production factors, such as water, labor, among others, and production as regressors (^{Bruneau, Renzetti, & Villeneuve, 2010}; ^{Féres, Reynaud, & Thomas, 2012}).

Finally, the third approach uses a production function, which relates the production value as a dependent variable and the monetary value or physical amounts of different production factors (^{Wang & Lall, 2002}; ^{Ku & Yoo, 2012}).

With regard to the latter approach, the methodology makes it possible to obtain the marginal productivity per production factor, corresponding to the implicit market price of the resource, being a greater contribution than the first two approaches. This situation is recommendable for study cases in which consumable prices do not necessarily reflect the true value, for example, as a result of setting public policies or market policies that benefit economic units, such as subsidiaries.

In the economic literature, the first two approaches are widely used. However, these approaches need establishing a water market, which translates into setting an appropriate fee structure and correctly implementing it, with data needed on costs and the marginal prices of consumables (^{De-Gispert, 2004}). Nevertheless, mean prices are used as a proxy for marginal prices. And as mentioned by ^{Dupont and Renzetti (2001)}, this translates into a deliberate bias, since perfect competition is when firms optimize marginal prices. On the other hand, a water market for industries depends on the available water source. Thus, for example, when this source is self-supplied, there is little or no external cost, whereas when it is supplied by the public network, the companies face an increasing block pricing structure (^{Renzetti & Dupont, 2002}).

Now, if market simplification is not possible, that is, mean prices cannot be used and the fee structure is unknown, or there is a fee structure but it is not properly implemented, the marginal productivity approach may be used to assess the marginal value of industrial water, since it does not need consumable prices. Accordingly, this research is based on this third approach.

There is less research on EVW for industrial use worldwide than for farming and household use. And this is almost inexistent on the national level. This is mainly because consumable prices, such as water used in the industry, in general, include subsidies and pricing policies do not totally reflect market prices.

In the studies where the third approach is used, ^{Wang and Lall (2002)} calculate the marginal value of water for a sample of approximately 2 000 companies in the industrial sector in China. They calculate a marginal value from USD 0.02/m³ for the electricity sector, up to a value of USD 11.5/m³ for transport equipment sector, with a national mean value equal to USD 1.1/m³ for the whole industry.

Meanwhile, ^{Aylward, Seely, Hartwell y Dengel (2010)} conducted a review of studies performed in different countries in Asia and North America, except Mexico, estimating that for the industrial sector the marginal productivity value of water presents a margin of USD 0.01/m³ to USD 6.94/m³.

Likewise, ^{Ku and Yoo (2012)} estimated the value of marginal productivity of water for the manufacturing industry in Korea, calculating a mean value for the whole industry equal to USD 3.7/m³, with the precision instruments industry having the lowest value and transport equipment industry the highest, USD 1.4/m³ and USD 44.3/m³ respectively.

Finally, ^{Vargas (2015)} calculated a marginal value equal to USD 30.9/m³ for the manufacturing industry in Chile, with a range between USD 11.8/m³ and USD 75.3/m³ for the textile and common metals sectors, respectively.

San Juan River Hydrological Sub-Region (HSRSJ)

HSRSJ Location

For administrative and preservation purposes of national waters, as of 1997, Mexico has been divided into 13 Hydrological-Administrative Regions (HAR). These HAR are comprised of groups of basins, considered the basic units for water resource management.

The borders of these HAR correspond to municipalities, in order to facilitate the generation of socioeconomic information.

The HSRSJ is part of the Bravo-Conchos Hydrological Region and the Rio Bravo Hydrological-Administrative Region (HAR Number 6) (Figure 1).

Figure 1 Hydrological Boundaries of the San Juan River Hydrological Sub-Region (HSRSJ). Source: Authors, 2017.  

HSRSJ has an area of approximately 30 847.26 km², covering municipalities in the states of Coahuila (4 municipalities), Nuevo Leon (37 municipalities), and Tamaulipas (8 municipalities). The main municipalities in this sub-region are Saltillo (in Coahuila), Monterrey, Guadalupe, and Apodaca (in Nuevo Leon), Reynosa and Matamoros (in Tamaulipas) due to its large economic activity and where the population is over 500 000 inhabitants.

HSRSJ is one of the most densely populated regions in Mexico, where around 7.2 million inhabitants are located (12.8% in Coahuila, 19.9% in Tamaulipas, and 67.3% in Nuevo Leon), with approximately 236 inhabitants per km², being one of the densest at the national level. At the same time, it occupies less than 3.3% of the national territory and, however, approximately 6.2% of the total population in the country is there, generating approximately 9.3% of the national GDP.

Methodology

Results

Estimation of Elasticities (ε) and the Marginal Value of Water (ρ)

The results of the estimation of the models for the entire manufacturing industry in the HSRSJ are shown in the following table. A total of 80% of variables in the Cobb-Douglas (CD) function are statistically significant from an individual point of view; while this value is 60% for the Trans-Logarithmic (TL) function in both models, with the majority of variables having a 99% significance level. The statistical adjustment (R ² ) in CD function was 90%, indicating that 90% of the total variation in the dependent variable is explained by the set of independent variables. In case of the TL function, the adjustment was 91%. Additionally, the value from the Fisher test (Prob > F), in both models, rejects the null hypothesis that all coefficients are simultaneously equal to zero (Table 3).

Table 3 Results of the Estimation of Econometric Models. Source: Authors, based on the 2014 Economic Census by the National Institute of Statistics and Geography (INEGI), 2017. 

Variable	Cobb-Douglas (CD)			Trans-Log (TL)
	Estimated	Significance		Estimated	Significance
lnL	0.012	0.614		-0.446	0.023	*
lnW	0.037	0.007	***	0.166	0.185
lnM	0.862	0.000	***	0.872	0.000	***
lnK	0.072	0.000	***	0.172	0.321
lnL*lnW				-0.011	0.570
lnL*lnM				0.092	0.000	***
lnL*lnK				-0.023	0.276
lnW*lnM				-0.044	0.000	***
lnW*lnK				0.039	0.003	***
lnM*lnK				-0.040	0.004	***
(lnL) 2				-0.055	0.004	***
(lnW) 2				0.001	0.869
(lnM) 2				0.017	0.003	***
(lnK) 2				0.009	0.366
A	1.425	0.000	***	1.195	0.214
N° Obs.	354			354
Prob > F	0.00			0.00
R 2	0.90			0.91

Statistical significance: * = 10%, ** = 5% and *** = 1%

The CD function found a water elasticity (ε_CD-W) in all manufacturing industry of 0.037, and a marginal value of water (ρ_CD-W) of USD 10.5 per m³ consumed and added to the production process.

For the TL function, the elasticity of the product with respect to water consumption for the entire manufacturing industry (ε_TL-W) was equal to 0.046, and ρ_TL-W was equal to USD 13.0 per m³. This was slightly higher than the findings from the CD function (Table 3). However, and as mentioned above, the literature reports a preference for the TL function over the CD because the latter is more restrictive than the TL function, which is a more general form of the production function (Table 4).

Table 4 Elasticity (ε_W) and the marginal value of the product with respect to water use (ρ_W) in the manufacturing sector Source: Authors, based on the Economic Census 2014 of the National Institute of Statistics and Geography (INEGI), 2017. 

Sectors	Code	Cobb-Douglas (CD)			Translog (TL)
		εCD-W (%)	ρCD-W (USD/m 3 )	εTL-W (%)	ρTL-W (USD/m 3 )
All manufacturing industries	31, 32 y 33	0.037	10.5	0.046	13.0
Food	311	0.027	9.7	0.024	9.1
Drinks and tobacco	312	0.071	14.1	0.068	13.1
Clothing	315	0.029	3.9	0.026	3.7
Leather	316	0.056	18.9	0.038	12.8
Paper	322	0.061	14.4	0.053	12.8
Chemistry	325	0.073	8.8	0.099	13.4
Non-metallic minerals	327	0.021	7.3	0.028	10.1
Metal products	332	0.009	1.3	0.053	7.3
Machinery and equipment	333	0.024	14.2	0.066	11.2
Computer equipment	334	0.024	0.4	0.047	0.9
Transportation equipment	336	0.082	15.9	0.075	15.4
Furniture	337	0.013	5.8	0.028	9.6
Other industries	339	0.083	11.1	0.073	8.5

Likewise, Table 4 shows the results of the different sectors comprising the manufacturing industry in HSRSJ, both for CD and TL.

The estimation of product elasticity with respect to water consumption ranges from 0.024 for the food sector, up to a value equal to 0.099 for chemical products sector. Meanwhile, the marginal value of water ranges from USD 0.9 per m³ for the computer equipment, communication, and blinds sector, up to USD 15.4 for the transport equipment sector.

When comparing the marginal value of water (ρ_W) to the percentage of the total value of production and total water use used in the production process, the most efficient sectors are transport, chemicals, and food (Figure 2).

Figure 2 Marginal value of the product with respect to the percentage of total production value and total water use in the manufacturing industry Source: Authors, 2017. 

Conclusions and recommendations

Despite the importance of the industrial use of water and the aspects that distinguish it from other types of use, the studies on the role of water in the industry are still very few. This is especially true for developing or emerging countries (^{Renzetti, 2002}). In the case of the San Juan River Hydrological Sub-Region (HSRSJ), the results from modeling the behavior of water used in the manufacturing industry confirm that there is sensitivity to economic variables. Specifically, it was estimated that the marginal value of water was equal to USD 13.0 per m³ used in the production process. At an international level, estimations of the same indicator are between USD 0.01 /m³ to USD 7.00 /m³ for countries in Asia and North America (excluding Mexico) (^{Aylward et al., 2010}), while in case of Chile, an average value of USD 30.9 m³ (^{Vargas, 2015}) was found for the manufacturing industry.

The value found in this research, although it is above average estimations at international level, can be explained mainly by the low costs related to the water used by the Economic Units in the country, and in this case the zone of study, which are below the prices of the water purchased by the EU in other countries.

The costs that EU should spend in the country, not only the manufacturing industry but all industry as a whole, are set by the “Federal Law of Fees - Applicable Provisions in Matters of National Waters 2016.” Those costs depend on the Availability Zone of the natural resource and whether that water comes from surface sources or subsoil.

The Availability Zone and underground water are classified by four Zones taking into account the Availability Index (Idas) which is calculated as Idas = Dma / (R - Dnc), where Dma is the annual mean availability of underground water in a hydro-geologic unit, R is the total annual mean recharge, and Dnc is the natural allocated discharge.

Based on this, Availability Zone 1 shows Idas lesser than or equal to -0.1 and therefore USD 1.12 /m³ should be spent in this zone. For Availability Zone 2, the Idas is greater than -0.1 and less than or equal to 0.1 and therefore USD 0.44 /m³ should be spent. For Availability Zone 3, the Idas is greater than 0.1 and less than or equal to 0.8 and therefore USD 0.14 /m³ should be spent. Finally, for Availability Zone 4, the Idas is greater than 0.8 and USD 0.10 /m³ should be spent.

In the case of HSRSJ, which belongs to Zone 1, the EU spent USD 1.12 /m³ on average, while, on average in the countries in the Organization for Economic Cooperation and Development (OECD) USD 1.38 /m³ (^{Walsh, Cusack, & O´Sullivan, 2016}) was spent. That is, a value below 21%, approximately.

The manufacturing industry uses water as a consumable for a variety of products; therefore, this resource is one of the most important consumables in the production process. Additionally, a steady water source is very important for manufacturing companies since it helps to avoid problems related to scarcity, and mainly to generate added value to its product. However, little research has been performed to analyze water in this industry and its added value in developing or emerging countries such as Mexico, or in this case in the HSRSJ.

In Mexico, the price of water is generally regulated by the State. Thus the importance of water and its added value in the end goods has not been considered in terms of the current dimensions.

This research estimated product elasticity (ε_W) and marginal value of water (ρ_W) for the manufacturing industrial sector located in the San Juan River Hydrological Sub-Region (HSRSJ) using information from nearly 18 000 economic units (EU), grouped at a municipal level through estimation of econometric models using the Cobb-Douglas (CD) and Trans-Logarithmic (TL) functions.

Product elasticity with respect to water for the manufacturing industry as a whole and the marginal value of water in the HSRSJ were estimated as 0.046 and USD 13.0 per m³ used in the production process, respectively. Estimated ρ_W ranged between USD 0.9 per m³ for the computer equipment sector up to USD 15.4 for the transport equipment sector, reflecting that this varies according to the analyzed sector.

From a public-policy point of view, the information found in this kind of research may serve to identify and quantify the monetary value of the public investment needed to offer water resources for the manufacturing sector or any subsector thereof. That is, the monetary benefit of water supply to the industry can be calculated as a result of a new project to provide this resource. For example, in case of the HSRSJ, on average each new m³ of water allocated to the manufacturing industry sector coming from new investment in the resource provision would result in an additional economic benefit of approximately USD 13.0 in the production process. Here, it is important that the cost per m³ of water incurred by the State or the company providing water service is lower than that amount, with the idea that such investment is feasible and cost-effective.

Likewise, the EU located in the HSRSJ spent an average of USD 1.01 per m³ used in the production process. Therefore, the State has a certain margin for considering an adjustment to the fees charged for water used for production processes, especially in basins showing water scarcity problems. This result makes it possible to reflect on the maneuvering room that the State has to increase prices for water supplied to the industry, including in this the costs related to damages to overexploited aquifers. This price modification will improve the use efficiency and resource location, increase State income and improve the government’s knowledge of water use.

In situations of water stress, as evidenced in HSRSJ, use efficiency shall be considered, for example, from the point of view of the added value offered by each subsector of the manufacturing sector in the production per m³ used in the process. For example, one m³ used in the food sector generates approximately USD 9.1 /m³ added value; while using the same m³ for chemicals generates USD 13.4; that is, 1.5 times more value. However, it is important to mention that this added value is completely from an economic point of view and does not consider effects for third parties, such as negative externalities, which may be several and to different degrees, depending on the subsector being analyzed and the economic agents involved. Accordingly, although the values found by estimating the marginal value of water in subsectors may be a useful tool for decision-making related to reallocating water use, especially when concerning water scarcity problems, it is necessary to complement this with studies on socio-economic assessments of the impact on environmental goods and services.

The information found by this research provides reasonable technical criteria for decision-making on water allocation and efficient water use. For example, in river basins with water scarcity problems this kind of analysis may serve to determine which subsectors, within the same manufacturing industry or other sectors (household, farming, among others) should have a certain degree of preference in terms of access to water, accompanying these studies with social and environmental considerations.

Likewise, these types of findings should make the society, authorities, and particularly the users of this natural resource become aware of considering the value they hold of water, scarcity, and the need of water, since it is considered an important resource in all production processes.

Therefore, we expect that estimations by these kinds of studies are useful enough to design future fee policies and water management policies by governments or administrations in charge of this scarce natural resource. It is also necessary to advance knowledge of industrial water use and generate instruments that help to achieve the rational management of water, in order to obtain better and greater benefits for society.

Finally, as a recommendation for future studies, environmental and social variables factors should be included in the specification of production functions (Cobb-Douglas and/or Trans-Logarithmic), such as: rainfall, recharge of aquifers from which companies exploit water, degree of exploitation of different aquifers, rigidity of consumables markets, labor strikes, among others.

Although these variables are not considered to be a direct consumable of the production process, they are variables that may be external to the process but that may have a certain degree of impact.

Furthermore, the findings by this kind of work may serve to measure how the production process economically impacts third parties, known in economics as negative externalities.