Changes in land ownership

Since the end of the 1960s, the Moche and Virú valleys have seen a series of changes in terms of land ownership, which have resulted in a new landscape for irrigation organizations. Before the 1970s, there were large privately-owned farms, and small, mostly poor, agricultural producers. With the Agrarian Reform of 1969-carried out by a military government-haciendas were expropriated and became agricultural cooperatives.

Subsequently, in the 1980s, cooperative properties were divided up throughout the country, except for the large sugar cooperatives; this transformed the Moche and Virú valley areas into small-scale production sites, that is, sites with smallholders and an average of approximately six hectares per family.

In the 1990s and 2000s, given changes in public policies that were strongly oriented towards the market and economic liberalization, such property was reconfigured through the process of buying and selling land, along with concessions of deserted land and the establishment of agro-export companies, mainly in the inter-valley territories. At the same time, small properties have survived in the Moche and Virú “old valleys” (^{Aste, 2018}; ^{Bury et al., 2013}).

The Chavimochic irrigation project

Within this context, the Chavimochic irrigation project was incorporated in 1994, with four valleys contributing to the effort: Chao, Virú, Moche and Chicama. The project consists of water diversion from one of the most important rivers along the coast, the Santa River-whose watershed belongs to the Ancash region (towards the La Libertad region’s valleys and inter-valley desert areas)-via the construction of a mother canal several kilometers long (see map in Annex 2). This allowed for the extension of the agricultural frontier into 46,700 new hectares, and the improvement of irrigation for 28,300 hectares; further, a permanent irrigation system was established in the area (^{Inga, 2016}; ^{Vos and Marshall, 2017}). The Chavimochic project’s work, from its beginnings in 1986 until its second stage, had a cost of US $960 million, which was financed by the Public Treasury (^{Inga, 2016}).

The project has allowed for an expanded and improved drinking water supply for families in the city of Trujillo, which has more than 900,000 inhabitants, in addition to urban and rural settlements within its zone of influence. However, it has to face the problem of the ongoing loss of glaciers in the Cordillera Blanca, which has the largest glacier coverage in the country, and which in the last 30 years has lost 34% of its surface due to the effects of climate change. These glaciers are the most important source for the Santa River’s water flow (^{Aste, 2018}; ^{Bury et al., 2013}).

It was a limitation of the present study, given its scope, not to address conflicts over water usage, especially between agricultural use and domestic use-or the conflicts between regions (Ancash and La Libertad) regarding the use of the Santa River water. It should be noted, however, that the Chavimochic project provided the Trujillo province with an important supply of drinking water, via the Alto Moche treatment plant (^{PECH, 2016}), which produces approximately 24 million m3 of water per year.

However, because irrigation water in the valleys under study was transformed from being seasonal to permanent, a result has been less use of and investment in groundwater wells, as costs have risen. The above circumstances, together with the permanent seepage of water into the subsoil, resulted in the area's aquifer groundwater levels rising until water emerged on the surface, creating in some places-especially in Virú-swamps and salinization of land. This problem resulted in public and private interventions promoting drainage projects and incentives to use groundwater.

Theories of collective action

Collective action refers to the cooperation of individuals to achieve a common goal. There are several theories that allow us to approach an understanding of this state of affairs; these are called theories of collective action. One such theory is explicated by ^{Olson (1965}), who, from a microeconomic approach, addresses the theoretical and practical difficulty of getting individuals with common interests to act or cooperate voluntarily in order to promote those interests and obtain the benefits that would result.

Another theory is that of ^{Ostrom (1990}, 2009). According to Ostrom, models that predict the failure of collective action (^{Hardin, 1968}; ^{Olson, 1965}) have limitations when it comes to explaining the facts, since they are based on extreme assumptions, rather than general theories and multiple methods that would allow for a closer view of the concrete reality and its complexities-especially regarding the management of common resources (^{Poteete et al., 2010}). In this sense, if the social context is one consisting of individuals who can defect but who are also capable of adopting cooperative norms, such individuals can be expected to establish contingent and appropriate commitments to follow rules which lead to the better functioning of social organizations (^{Agrawal, 1996}; Ostrom, 1990; ^{Wade, 1988}).

A comparative study by ^{Agrawal (2001)} examines three of the most influential works on common resources and their implications for collective action, these being the works of ^{Baland and Platteau (1996)}, ^{Wade (1988)} and ^{Ostrom (1990}). He does not find unanimity among them regarding a set of relevant variables that explain the sustainability of institutions that shape the use and management of common resources. However, he finds that the approaches coincide in terms of the importance they give to developing institutional agreements which support locally-designed resources management and access regulations, and the need for easy application of these regulations or ways to enforce compliance. Agrawal also finds that-among a large number of variables-group size is an important one in explaining the possibilities of sustainable collective action; another key variable he identifies is having clearly defined group boundaries.

Furthermore, several case studies (^{Araral, 2009}; ^{Cody et al., 2015}; ^{Fischer and Qaim, 2014}; ^{Gardner et al., 1990}; ^{Putterman, 1983}; ^{Wade, 1987}, ¹⁹⁸⁸) demonstrate that achieving collective action is more likely when the size of the group, or total number of decision makers, is small. That is, if the group is small enough for people to be able to interact with each regularly, collective action will be more easily sustained. However, there is debate on this issue, as other scholars conclude that the size of the group is not significant, or at least that the degree of sustainability of the collective action is uncertain (^{Nagrah et al., 2016}), and give greater importance to other variables such as the heterogeneity of the group (^{Ruttan, 2008}; ^{Villamayor-Tomas, 2014}).

Other environmental variables, such as favorable prices or increasing demand for local products, can strengthen common interests, which makes it easier for the group to agree on collective action (^{Agrawal and Goyal, 2001}; ^{Fischer and Qaim, 2014}; ^{Ito, 2012}). On the other hand, theories of collective action can also explain the emergence of restrictive institutions or rules that limit individual behavior in favor of obtaining collective benefits (^{Chang and Evans, 2007}), as well as the role of trust and positive leadership in the groups that make collective action more likely to be sustained (^{Nagrah et al., 2016}).

Collective action and irrigation water management

The sustainable management of irrigation water is a complex and vital issue for farmers. In Peru, User Boards (organized by valley) that belong to a common hydraulic sector are in charge of this management; these are made up of User Commissions (organization of irrigators according to irrigation branch line) based on a hydraulic sub-sector. Within this structure, the organization’s larger group is constituted on the basis of smaller user groups, which enables cooperation to develop from the ground up (^{MINAGRI, 2015}; ^{Muñoz, 2009}). Several international case studies suggest that this structure lowers the costs of organization, and promotes collective action towards the management of common resources (^{Nagrah et al., 2016}; ^{Ostrom, 1990}; ^{Wade, 1988}).

The User Boards are responsible for planning the distribution of irrigation water at the level of the entire valley to which they pertain, in addition to collecting water rates and representing all of the users in the corresponding irrigation district. The User Commissions are given the power to execute and control the distribution and delivery of water within their jurisdiction, as well how to establish penalties for violators of the organization's norms (^{MINAGRI, 2015}; ^{Muñoz, 2009}).

The social irrigation organizations in each of the valleys under study are: the Water User Board of the Moche River Basin (WCBMRB), made up of 4,321 irrigators, and the Water User Board of the Virú River Basin (WCBVRB), with 2,596 irrigators. Each board is made up of User Commissions that are the basis of the irrigation organization. The WCBMRB is made up of 11 User Commissions, where the smallest group is made up of 157 irrigators, and the largest of 780; the WCBVRB is composed of 13 User Commissions, with 33 irrigators in the smallest group and the largest group made up of 586 irrigators.

Additionally, the Pressurized Irrigation Users Board of the Moche, Virú and Chao Irrigation District has been created to facilitate the distribution of the Chavimochic canal’s irrigation water. This board is mainly made up of agribusiness companies connected to the export sector, with a total of 55 users. These companies are mainly located in the inter-valley zones, where the land reclaimed from the desert by the Chavimochic project are located. The project diverts water from the Santa River towards the Moche, Virú and Chao “old valleys” and inter-valleys (^{Bury et al., 2013}; ^{Vos and Marshall, 2017}).

It should be noted that the Peruvian State provides legal support for collective management of irrigation water, as laid out in the General Water Law, Executive Decree 17752 of 1969, and regulated by Supreme Decree 037-89-AG of 1989, which stipulated the type of organization that users should form and gave them the capacity to establish and collect irrigation water rates. This function was ratified by the Water Resources Act of 2009. These legal mechanisms, which are typical of state action, seem to have played a fundamental role in strengthening collective action around water management.

Moreover, taking into account the importance of group size in the development of cooperation, there is evidently a diverse range of group sizes between the Moche and Virú User Commissions. The theory referred to above indicates that if the group is small enough for people to interact with each other several times, then there is more of a possibility that they will cooperate-since actors can negotiate and agree more easily on their common goals and proposed actions (^{Marshall, 2014}; ^{Muñoz, 2009}).

In these valleys, changes or variation in local prices, as well as in agricultural production, are factors that should have an effect on farmers’ collective action. If both variables move in a positive direction, farmers are incentivized to seek expanded production and demand, and therefore, require irrigation water in greater quantities. The foregoing means that the total amount from the water rate being collected is higher, and therefore, the outcomes of collective action are strengthened by virtue of the groups having more resources available to cover the work of irrigation infrastructure operation and maintenance, as well as to run the user organization.

Research question and hypothesis

The question that motivated this research project was: how do group size, changes in prices, value of local production and concentration of land make collective action more sustainable, stronger, or weaker in the Moche and Virú river valleys located on the north coast of Peru?

The hypothesis of the present study can be stated as follows: collective action by irrigator or user groups-as expressed in the economic result or efficiency of irrigation water rate collection, and defining said result as the gap between expected revenue and revenue collected-is more likely to be sustained with smaller group size, positive movement both of prices of goods produced in the valley and growth of local agricultural production value, and with a lower concentration of land that comprises each user commission in the valleys under study.

It is postulated that the gap in the water rate collected by the User Boards depends on the number of members that make up the User Commissions and the conditions of the economic context in which they carry out their productive activity (price changes and agricultural production), in addition to the specifications of irrigated land concentration in the User Commissions.

Collection of data and information

Field visits were made up to three times to obtain data and statistics on the Moche and Virú User Boards, which are boards that have recorded and systematically tracked information about irrigation water rates since 2000. Data were also collected regarding the number of irrigators and the amount of land which was the property of or owned by each user commission.

Data for prices of the primary products of both valleys, and the gross value of agricultural production, were taken from the National Institute of Statistics and Informatics [Spanish acronym INEI] and the Ministry of Agriculture and Irrigation (MINAGRI). The calculations for price changes and the growth of agricultural production gross value are prepared by the author.

It is worth highlighting the way in which irrigation water is delivered to most of the irrigators in both valleys. User Boards deliver the water to the irrigators after they have paid the value of the volume or quantity of water requested. For a user who has not become part of the new system (because of being located on higher ground, or other reasons), the criterion of a fee paid to the User Commission is maintained, so that they do not have to be in default to receive their share of water. The new system is implemented on delivery, meaning that there is no possibility to have water on the property without first paying for it. Each User Commission is responsible for collecting the water rate within the new system, under which each user must make an advance payment for the volume of irrigation water required for their crops. To do this they have constructed infrastructure in the fields and hired personnel for the job. The money collected comes to the User Boards, which are responsible for delivering water to the irrigators and administering the money received, and for the operation and maintenance of the irrigation system.

This system counteracts the possibility of late payment of the service fee, and has been made possible due to the new, modern infrastructure built along with the Chavimochic project that serves to redirect water from the Santa River to the Chao, Virú and Moche valleys. To highlight the importance of this project for the region, it can be noted that the total agricultural water demand in this area is 596 hm3, 87% being covered by the diversion of waters from the Santa River (^{Aste, 2018}). The rest is made up of water from the valley rivers, lagoons and aquifers in the area.

Variables

The following variables were established based on the literature reviewed above, especially the variables that could be utilized by working with the data that was collected, or that would allow the construction of an indicator suitable to test the hypothesis regarding the outcomes of collective action.

Dependent variable

Independent variables

Model and calculation method

The methodological strategy of this study is based on the characteristics of the hypothesis that we seek to prove empirically. This hypothesis arises from the theory of collective action which assumes that group size is a fundamental factor in sustaining group cooperation. For the Moche and Virú irrigator groups, or User Commissions, cross-sectional data are available for a period of eight years. A data panel is available which can be used to perform the regressions that allow for a comparison with the stated hypothesis. The random effects panel model method was chosen to calculate the variables, allowing the unobservable heterogeneity of the units of analysis to be captured.

The Users Commission was taken as a unit of analysis. The variables to be used for the econometric model are described below, and the equation for the corresponding regression is as follows:

Where:

Y _i,t : Gap (revenue received minus expected revenue) in irrigation water rate collection for each User Commission in the year corresponding to the 2004-2011 period. This is the dependent variable.

X _i,t : Number of irrigators in each of the User Commissions for each year corresponding to the study period. This is the independent variable.

Z _t-1: Growth rate of agricultural production gross value in the La Libertad region, in each of the years of the study period and using a lag of one year. This is another independent variable, a proxy for the economic-productive environment.

J _t-1: Inflation rate of the Moche and Virú valleys’ primary products. This is the inflation rate created for the years corresponding to the study period, with a lag of one year. It is another independent variable, and another proxy for the economic environment.

I _t : The Herfindahl-Hirschman Index (HHI), which measures the degree of concentration for irrigated land in the Moche and Virú valley user commissions.

α: Intercept.

β₁, β₂, β₃ and β₄: Regression coefficients.

u _i,t : Error term between groups.

𝜀 _i,t : Error term within groups.

Calculations for the Moche River Basin User Board

The results of the regression for the Moche Board are presented in Table 1.

The first result is the average variation of Y (gap) when X (number of irrigators) varies throughout the groups in 1 unit. In this case, the ratio is negative and 99% significant. That is, Y widens, forming the most negative gap at 0.29815 units when X increases by 1 unit throughout all of the groups.

Accordingly, it can be noted that the number of irrigators variable is significant in the explanation of the gap between received and expected revenue. This indicates that the addition of an irrigator (user) in a commission could cause the gap to increase by 0.29815 monetary units. This mechanism results in the decrease of revenue received by each Irrigation Commission, and means that the amount collected moves away from the expected or target revenue.

An interpretation of this result would be that with a small number of irrigators it is possible to agree on a higher irrigation water rate than with a large number of irrigators; this is due to the fact that it is more complex and difficult to negotiate the tariff amount within a Commission if a large number of irrigators are involved in the decision, all of whom have to be heard and negotiated with, and who typically seek to reduce the rate proposed by the Board of Users’ technical body. In turn, there would also be less revenue, since the amount depends on the rate established for the corresponding year, and thereby the gap would also increase.

The second result, the growth rate of agricultural GVP in the region, is not significant. The third result, the local inflation rate of the valley’s primary products, is 99% significant. This result indicates that the average variation of Y (gap) is reduced when the inflation rate varies (rises) by 1 percentage point; in this case the relationship is positive. That is, the gap becomes less negative at 121.0155 units when J (price index) increases by 1 point. Subsequently, for each percentage point that the inflation rate rises, the gap is reduced by 121 monetary units.

This result would indicate that an increase in the prices of products sold by the valley’s producers encourages higher production of said products, therefore demanding a larger amount of irrigation water from irrigators or producer-users. Additionally, this allows the unit water rate to be sustained at levels that generate greater revenue.

In the fourth result, the HHI is not significant. The R-squared (R-sq) shows how much Y’s variance is explained by the variables included in the model. The R-sq (overall) is included in order to demonstrate the adjustment index at the level of the model as a whole, which is 0.5773.

Likewise, three scenarios are presented for the Moche Board, examining the relationships between the model’s dependent variable and its independent or explanatory variables (see Table 2).

The first scenario is the one analyzed above, in which the variables corresponding to number of irrigators and local inflation rate are 99% significant. In the second scenario, in which the HHI is removed from the regression, the same variables as in the first scenario continue to be significant, also at 99%. The local inflation rate is withdrawn in the third scenario, where only the irrigator number variable emerges as 99% significant in the explanation of the gap.

Calculations for the Virú river basin User Board

The results of the Virú Board regression are presented in Table 3.

The result for the irrigator number variable-in the explanation of the gap between received and expected revenue-is not significant. In this case, the theory of collective action which postulates that the number of members is fundamental for the development of cooperation is not corroborated by empirical evidence.

The result for the growth rate of agricultural GVP in the region is 95% significant in explaining the gap. This means that one point of growth in the region’s agricultural production allows the gap to decrease by 1.06375 monetary units. Given that a lag of one year is used, this impacts the decisions that producers make in the following year, when their expectation will be to increase production and, therefore, demand more water-which increases the revenue received by the Board.

The result for the local inflation rate is 99% significant. It indicates that the gap is reduced when the local inflation rate varies (increases) by 1 percentage point. The relationship in this case is positive-that is, the gap is reduced by 77.7195 monetary units when the local inflation rate increases by 1 point. This result is interpreted to mean that an increase in the price of agricultural goods sold by producers encourages greater production of said products, and consequently leads to a higher demand for irrigation water. This allows the unit water rate to be sustained at levels which generate more revenue for the Board.

The result for the HHI is 95% significant in explaining the Virú Board’s gap. In terms of the gap, the relationship is negative, a fact which can be interpreted to mean that a higher concentration of irrigated land grants more negotiation power to the irrigation commissions, such that the rate is lower-as postulated in the specifications of the model.

The R-squared (R-sq) that is reported is the R-sq (overall), showing the adjustment index at the level of the entire model, which is 0.4196.

Likewise, for the Virú Board, three scenarios of the relationships between the model’s dependent variable and independent or explanatory variables are presented (see Table 4).

The first scenario has already been analyzed, in which the local inflation rate variable is 99% significant, the agricultural GVP growth and HHI variables are 95% significant, and these show a relationship to the dependent variable, as described in the analysis above. In the second scenario, in which the HHI was removed from the regression, the local inflation rate continues to be 99% significant; however, here the number of irrigators and the growth of agricultural GVP are also significant, at 90% and 95% respectively. In the third scenario the local inflation rate was removed, and the HHI is found to be 95% significant, while the growth of agricultural GVP is significant to 90%, and the number of irrigators is not significant in explaining the gap.