1Water pollution from agricultural sources has been a persistent problem in the region of Brittany [Bretagne] since the 1970s. Officially put on the political agenda only in the early 1990s, [1] the issue has since prompted a growing number of public policy responses – in the form of both voluntary-based and binding programmes. In the course of studying these programmes devised to fight agricultural pollution and “restore water quality” [reconquête de la qualité des eaux], I have addressed the impact of this environmental problem on agricultural public policy, and by extension the question of change in public policy, traditionally a major concern in public policy analysis. There are different schools of thought on policy change, including policy learning theory [2] and the rational choice approach. [3] The neocorporatist theories that have long prevailed in the agricultural sector have mostly served to evidence the existence of a phenomenon of inertia in lieu of expected change, emphasizing the influence of agricultural professionals and the close ties some of them have managed to develop with the Ministry of Agriculture. [4] Other particularly fashionable political science theories have described instances of continuity of public policy, emphasizing the role played by institutions – by “institutions”, these authors refer to the standards and the rules of the game that regulate the actions of individuals. [5] In other words, they argue that long-established public policy habits and procedures stand in the way of objectives or devices newly introduced by political actors. Many neo-institutionalist authors stress the tendency of these institutions to foster relative continuity rather than change in public policy. They do admit that some change occurs, but according to them it happens incrementally [6] or it is brutally imposed from the outside. [7] However, other authors argue that it is important to consider continuity and change within institutions together. [8]

2In this article, I follow in the footsteps of the neo-institutionalists by setting out to evidence the ability of institutions to resist pressure to change. I show that this resistance is not primarily the consequence of the mobilization of agricultural interest groups against a reform of the intensive agricultural production system; rather, it is due to the institutions themselves. I demonstrate that local agricultural policies resist the change brought about by water quality restoration measures as a result of the existence of instruments liable to paralyze change. By “instruments” I mean the “technical public policy instruments” defined by Pierre Lascoumes and Patrick Le Galès, i.e., devices, regulation techniques and methods of operation in public policy. [9] According to these authors, “instruments are institutions in the sociological sense[:] a more or less coordinated set of rules and procedures that governs the interactions and behaviors of actors and organizations”. [10]

3I consider instruments as fully-fledged institutions with a view towards building bridges between neo-institutionalist theory and the policy instrumentation approach. I intend, for instance, to discuss the application of some of the concepts used in various strands of neoinstitutionalist thought to the policy instrumentation approach. I will specifically examine the instruments’ cultural footprint in light of the sociological neo-institutionalist view of instruments as first and foremost the expression of a shared worldview. I will also discuss the rational-choice neo-institutionalist view of institutions as facilitators of social exchanges, [11] as well as the historical neo-institutionalist concept of “path dependence” – regarding the latter, this paper will highlight a new rationale of institutional dependence.

4In path dependence theory, the recurring use of norms or institutions is thought to trigger a mechanism of “increasing returns”. [12] Here, I show that dependence can also come from the institutions themselves, and in the case at hand from the instruments used more than from the path taken by the institutions. To sustain my argument, I show that the univocal view of institutions as facilitators of exchanges, tracing a path for public policy that is easier to follow than to avoid, is an erroneous one. I show that dependence on public policy instruments, those specific types of institutions, results from the cumulative effect of the ideological footprint on these instruments, leading to a regulatory imbroglio that ultimately causes a movement of distancing between these instruments and public decision-makers. In the case under study, institutional dependence is actually rooted in complexity, rather than in increasing returns measured in terms of coordination of social interactions.

5First, I will follow the approach of Lascoumes and Le Galès, according to whom “a public policy instrument constitutes a device that is both technical and social, that organizes specific social relations between the state and those it is addressed to, according to the representations and meanings it carries”. [13] In keeping with sociological neo-institutionalism, I show the extent to which these public policy instruments are shaped by ideological representations that have little to do with the new environmental objectives imposed from the outside by new actors in the agricultural field, and are rather informed by the values invested since the 1960s by the promoters of these instruments in departmental [14] agricultural public policy. I demonstrate that despite the external imposition of a new environmental frame of reference in agricultural public policy, the instruments set up to achieve the new environmental objectives are still as indebted, if not more, to the productivist frame of reference [15] that has prevailed in Brittany’s agricultural circles since the 1960s.

6Secondly, I will shed light on a less extensively developed dimension of Lascoumes and Le Galès’s work on policy instrumentation. The authors mention an effect of autonomization of the institutions and instruments created: “as they are used, [instruments] tend to produce original and sometimes unexpected effects”. [16] I provide a concrete illustration of this point in the second and third part of this article. Here, the continuity of public policy is not explained by the influence of the past on the present (historical neo-institutionalism) or the perspective of high costs (of coordination, learning) [17] in the event of a change (rational-choice neo-institutionalism), but instead by the complexity introduced by these instruments and the dependence they create for the decision-makers. In the case discussed, rather than creating a dependence on the path laid out, institutions create dependence on the instruments that map out change and/or stability in public policy.

An inadequate response: the effect of sociological neo-institutionalism on departmental regulation

7In Brittany, new environmental objectives in the agricultural sector and particularly in intensive pig and poultry farming were introduced as a result of several local and European factors. I will mostly focus on the European level here, as the EU has imposed particularly stringent new standards on factory farming since the early 1990s. By way of comparison, over the same period, the pressure of local actors committed to change in agricultural practices has only resulted in the adoption of an incentive-based programme, the national programme for agricultural pollution control [PMPOA, Programme national de maîtrise des pollutions d’origine agricole].

8The EU has been issuing regulation on protection of water resources since the 1970s. In 1991, the “Nitrates directive” [18] established a maximum standard of application of livestock manure on farmland. [19] As for any other directive, member states were left entirely free to choose how they intended to ensure compliance with the new standard. Each of them therefore had to specify a method for managing the colossal surplus of livestock effluent created by this new limit in intensive farming countries.

9Adapting this new, EU-imposed environmental objective to Breton agricultural actors long used to a productivist frame of reference raised a major challenge. In the battery farming system developed from the mid 1960s in Brittany for pig and poultry farming, production does not depend on land availability. The system is said to be intensive because it makes it possible to breed many more animals than the overall surface of farmland effectively available would normally allow. Hence, there is no dependence on the forage and food produced by the land for the animals bred on site. With the modernization of farming, food and forage could now be imported by the breeders, therefore making it possible for regions such as Brittany, where the soil is not very fertile and the land-to-farmer ratio is small, to produce in spite of unfavorable natural conditions. Yet, with the agricultural pollution problem, the physical constraints farmers had tried to escape upstream of the production process now reappeared downstream: the waste from livestock farms cannot be eliminated in acreages too small to absorb it. This is a so-called structural problem, [20] as precisely because of the farming model promoted at regional level, crop fertilization cannot fail to be unbalanced and pollution of water and soil is almost inevitable. [21]

10Central ministerial administrations, working closely with decentralized branches of the Ministry of Agriculture, developed action programmes under the Nitrates directive in which numerous provisions seem superfluous and surprisingly complex, particularly in an area already characterized by great complexity. [22] The methods used to achieve compliance with EU standards are rooted in a solidarity-based approach to the livestock manure problem in France: they decreed and organized solidarity between farmers. The emphasis on solidarity is all the more striking when contrasted to the Dutch system set up in response to similar issues, which remained faithful to a more individual, farm-by-farm approach. [23] A number of public policy instruments such as the “manure application ceiling” [plafond d’épandage] and the “mandatory treatment threshold” [seuil d’obligation de traitement], which I discuss later in this paper, attest to the solidarity-based approach chosen in France. This choice can be explained by the prevalence of ideological schemes that the Ministry of Agriculture departmental office [DDA, Direction départementale agricole] and the majority farmers’ union have come to share over time, and by their close collaboration within the framework of agricultural co-management [cogestion agricole], [24] as we will see. [25]

A discriminatory device

18The FDSEA now claimed to oppose the implementation of a plan it had originally conceived as voluntary-based, transformed by the DDA into a set of binding regulations. [35]As it was also eager to refrain from speeding up the already ongoing process of the disappearance of small farms as ever-more intensive methods of production were being used, the DDA devised a particularly interventionist plan. This amount of interference was all the more surprising as the problem and therefore its solutions appeared particularly sensitive from the start. [36] The DDA thus took the risk of further increasing their complexity. This approach received support from the head offices of the Ministries of Agriculture and Environment, since the Voynet-Le Pensec circular adopted in 1998, which presents the French plan for implementing the Nitrates directive, endorsed the administrative categories created locally. [37] On agricultural pollution in general, Breton departments, and particularly the Côtes-d’Armor DDA, were forerunners of national policy – first because Brittany was the first region to face this problem in the 1980s, and second because the then Minister of Agriculture in the Jospin cabinet, Louis Le Pensec, was a former socialist Breton MP. [38] Le Pensec had also served as vice-president of the Finistère General Council when the Confédération paysanne was for the first time at the helm of the departmental Chamber of Agriculture: he was also convinced of the need to defend small livestock farms.

19Three more years passed before the second round of prefectoral bylaws on the implementation of the Nitrates directive were issued and the plan was finally made operational. Until then, only polluted areas ( “structural surplus areas” – in French zones d’excédents structurels) had been identified; no farms had been singled out. With the 2001 bylaws, the DDA now faced the task of determining who among the population of surplus-producing farmers was going to be targeted. The DDA did not actually stop there and went as far as to specify categories among those farmers: those eligible for manure application and those forced to invest in a treatment facility. To calibrate this plan, they adopted a global approach by structural surplus area. First, at departmental level, they calculated the global amount of surplus to absorb for surplus-producing farms [39] and the surface of land potentially available for manure application in deficit farms. On that basis, they determined what they called a “manure application ceiling” [plafond d’épandage] for each structural surplus area in the department (in the form of a maximum number of hectares made available to a farm for manure application) and a “mandatory treatment threshold” [seuil d’obligation de traitement] (once the threshold has been reached, the effluent that cannot be applied within the manure application ceiling will have to be treated). With the system put in place, as soon as a surplus-producing farm reaches a certain level of effluent production, it automatically falls into the category of mandatory treatment and becomes ineligible for application on farmland owned by deficit third parties. Conversely, other surplus-producing farms whose production levels remain under that threshold are allowed to apply manure on land loaned by third parties. The DDA’s ambition here is to prevent bigger farms from having massive recourse to land application, when they are perceived as the most financially sound and therefore the most suitable candidates for investment in treatment.

20Ultimately, this plan seems unnecessarily complex in many respects, because it is very interventionist and partly discriminatory, without this discrimination being related to the environmental issue at hand. The plan can be better understood if we consider its strong social dimension, beyond the environmental aspect. It also attests to the state of tension pertaining to land access issues, always a sensitive subject in the Côtes-d’Armor department. Long-established positions shared by the administration and the majority farmers’ union sometimes even led public authorities to let the environmental situation worsen for social reasons.

Instruments and their negative side effects

25These instruments however soon got out of hand and beyond the control of public decision-makers. The multiple provisions that had been intentionally and carefully introduced ended up creating a bureaucratic labyrinth, producing numerous negative side effects that undermined the balance of the regulation, as we will see. Here I wish to emphasize an aspect that strongly conflicts with the utilitarian explanation of path dependence through a mechanism of “increasing returns”. In path dependence theory, institutions endure because they facilitate coordination and produce learning effects in social interactions. The case under study appears to contradict this: here, instruments – institutions – do not necessarily facilitate coordination between actors or allow them to grasp the complexity of the situation. Quite the opposite in fact: they make things more complex and reduce the possibilities for actors to learn change.

Understanding (the lack of) change through the analysis of instruments

44As environmental organizations keep pointing out the absence of results or improvements in terms of water pollution by nitrates, there has been more and more pressure to come up with a new way of addressing the problem. Ever the outsiders of manure management policy, these organizations have tirelessly used all the resources available to them to further their cause. This includes numerous legal appeals, both at national and EU level. In one of these cases, brought in 1992, the European Court of Justice ruled against the French government in March 2001 and June 2007. The European Commission had opened infringement proceedings for failure to comply with a legal provision of the 1975 Surface Water directive. [77] Since the Maastricht Treaty, the Commission was entitled to impose financial sanctions to pressure France into complying with the law as quickly as possible. This threat was taken very seriously at the highest levels of French government, as the Commission had carried out similar threats for the first time in July 2005 in the so-called “undersized fish” case, which cost France 77 million euros overall. Since that case, the Secretariat General for European Affairs, reporting directly to the Prime Minister’s office, monitored all EU litigation where article 228 of the Maastricht Treaty could apply. [78] Other Ministries were also on their guard, since for various reasons, the Ministries of Transport, Defense and Justice, among others, had to share the burden of the fine for authorizing the catching and selling of undersized fish with the Ministry of Agriculture – a lump sum of 20 million euros and a daily penalty payment of 316 000 euros. In the “surface water” case, the penalty payment was lower – 117 000 euros daily – but even the most optimistic studies forecast that compliance with EU standards in problematic catchment areas will not be achieved before 2015. [79] Understandably, then, the other Ministries pressured the Ministry of Agriculture into acting quickly. Yet, while it was now necessary to devise an emergency plan at national level to convince the Commission not to use financial sanctions, the departmental plan remained relatively stable even though it had yet to produce any tangible effects on nitrate concentrations in watercourses – which was problematic in the light of the Surface Water directive. To respond to European Commission pressure, instead of reworking a plan that had proved impractical, regional authorities made it even more complex by creating yet another new administrative category – “catchment areas targeted by EU litigation” [bassins versants en contentieux européen] – to be monitored more closely, with financial incentives for farmers to shift towards extensive or organic production. The French plan also included manure application restrictions more stringent than those set out by the Nitrates directive, [80] but only in the four catchment areas deemed the most problematic, failing to anticipate likely similar trends in other catchment areas. The authorities also deliberately ignored the fact that many of the studies they had commissioned had expressed skepticism regarding the efficiency of the plan. [81] Ultimately, these measures were merely added on top of the existing manure management plan, which remained intact despite its failure to produce results.

45Here, I argue that this continuity in policy is not the result of path dependence as described by historical neo-institutionalists. In the case under study, continuity in policy is not due to the fact that policy settings inherited from the past are perceived as more adequate or less costly. Rather than path dependence, the manure management case is one where dependence on instruments prevails. This dependence does not exist because these instruments make public policy implementation more practical for the authorities. As we have seen, no mechanism of “increasing returns” can be observed here, but rather the contrary. Conversely, it is because they make public policy more complex, as opposed to easier, that these instruments become obligatory passage points of sorts; they function as anchors in the chaotic environment they have contributed to create or foster. In other words, I argue, as an alternative to historical neo-institutionalism and its path dependence concept, that the authorities stick with these instruments in order to contain complexity rather than because they simplify matters in the first place. Regarding the issue of water pollution from agricultural sources, they steer a short-sighted course, sticking to the instruments as much as possible, without the latter – due to their very complexity – making it possible for them to easily follow an existing path.

46This dependence on instruments becomes clear when we observe the way the majority agricultural coalition has become increasingly looser and even fragmented over time. [82] I argue that the stability of Côtes-d’Armor policy cannot be explained by the maintenance of shared cognitive frames within the agricultural coalition. First, the DDA recruited new staff members whose backgrounds were less close to those of agricultural professionals. The civil servants in charge of the first manure management programmes came from Brittany and were trained at the School for agronomic studies in the regional capital Rennes [ENSA, École nationale supérieure d’agronomie]. But as of 2003, the new head of service in charge of agricultural regulation had no local ties, and had trained at the Paris-based Ecole Polytechnique and National school of rural water and forestry engineering [ENGREF, Ecole nationale du génie rural, des eaux et des forêts]. He had fewer ties to the agricultural world than his predecessors. Additionally, members of the majority coalition became increasingly distant over time, as the types of exchanges between members of the administration and members of the majority farmers’ union changed: in the 2000s, there was a gradual shift from technical and generally consensual consultations towards more negotiation. [83] Over the same period, discussion was relocated from the departmental level, which used to be the traditional setting of agricultural co-management, to the regional level with the creation of the MIRE [84] in February 2001, reporting to the Secretariat General for Regional Affairs, i.e., the regional prefecture.

47As a result of these changes, interactions between the majority agricultural organizations and the agricultural authorities became more politicized. Agricultural representatives had no qualms about enlisting local and national political supporters to pressure their ex-partners in the DDA. They now influenced public policy in a political manner, using local MPs to reach the prefect and his decentralized head of service, whereas under the technocratic decision-making system that prevailed from the 1960s in the agricultural sector, influence tended to work the other way around, from the director of the DDA, an unfailing ally of the majority farmers’ union, to the prefect, the arbiter and mediator on local interests. The ideological proximity of the agricultural coalition members, which explained the recourse to the specific instruments of the departmental manure management plan, cannot explain their stability over time.

48The challenges faced by lobbying operations of the majority farmers’ union are another illustration of the constraint exerted by the instruments themselves. At first sight, this lobbying appears to have paid off at local level, as the prefect granted several requests made by agricultural economic actors. In particular, agricultural cooperatives had been eager to restructure livestock farms for a long time, by allowing farms to expand by purchasing other existing farms. This was in theory impossible since the 1990s, as extending livestock farms was forbidden in surplus areas. Yet, the authorities ended up conceding this possibility in 2005, considering it did not affect global production volumes at departmental level. [85] As it led to further development of intensive farming and a concomitant decrease in the number of active farmers, environmental organizations and the Confédération paysanne have always opposed this measure. [86] In their own words, they denounced, first, the breakneck quest for productivity, and second, the risk of further worsening the environmental situation of the catchment areas affected by this change – the catchment area being a more relevant unit than the department to assess the consequences of a restructuring. The adoption of this measure ultimately led these two interest groups to disengage from the debate on regional agricultural policy reform launched in the early 2000s with the “Charter for sustainable agriculture in Brittany” [Charte pour une agriculture pérenne en Bretagne]. Yet, if the regional prefecture appears ready to make major concessions to regional economic interests, this does not mean that the authorities will grant any kind of request from the majority farmers’ union. Due to the features of the instruments in place, their lobbying has limitations. We may for instance consider the fact that during several years, the majority farmers’ union asked for a revision of manure management objectives, on the legitimate grounds that numerous poultry and cattle farms had gone out of business as a result of multiple economic crises since the early 2000s, when the manure management objectives were established. Nevertheless, with the aim of containing the complexity of the regulation, the regional prefecture has always rejected this request; the objectives had to remain intact. In a context of relative uncertainty on the balance of the manure management plan, these objectives are somewhat understandably held as non-amendable.

49Ultimately, I have endeavored to show the unexpected consequences of the use of instruments on public policy and the way instruments become emancipated from their producers in practice, in an illustration of the specific dynamic of policy instruments evidenced by Lascoumes and Le Galès. According to these two authors, instruments “generate their own effects, beyond the intentions of decision-makers”. [87] Other authors also described as a “ratchet effect” the irreversibility of the choices made and the relative powerlessness of actors towards the instruments. [88]

50* * *

51In this article, I have addressed the concept of change in agricultural public policy in the Côtes-d’Armor department. I have shown that despite the introduction of new environmental standards and objectives, particularly by the EU, all perspectives for change are nipped in the bud by the enduring traditional ideological schemes deriving from the productivist paradigm that has prevailed since the 1960s, which inform the public policy instruments that are supposed to help achieve these objectives. More specifically, the negative side effects of these instruments in their practical use attest to their ability to neutralize opportunities for change and in the process strip the new objectives of their substance, including when the persistence of the environmental problem led the agricultural coalition to become more fragmented over time.

52Within a neo-institutionalist perspective, I have focused more closely on the study of public policy instrumentation first, to show the impact of ideological schemes, as instruments are not axiologically neutral, “on the contrary […] they are fueled by one interpretation of the social and by precise notions of the mode of regulation envisaged”; [89] second, to gauge the progressive emancipation of these instruments from the sociological institutions to which they were initially supposed to be faithful. In this sense, this article is a concrete illustration of Lascoumes’ claim that instruments sometimes end up beyond the control of actors, who become prisoners and somehow dependent on the tools they have forged.

53Although the path dependence phenomenon is usually explained in terms of a utilitarian rationale of “increasing returns”, [90] I have suggested an alternative interpretation of this dependence. The instruments of information used and the administrative categories consequently created induced a very specific public policy course which, quite strikingly, instead of simplifying matters, brought even more uncertainty on an agronomic issue that was very complex to begin with.

54While the utilitarian interpretation, in line with Arthur, allows for “a potential path inefficiency – in the long-run, the outcome that becomes locked in may generate lower pay-offs than an alternative solution would have”, [91] according to Arthur increasing return mechanisms are mainly triggered by learning effects and coordination effects. In the case under study here, we have observed that there were no learning effects but instead much confusion surrounding these institutions. The endurance of institutions due to the high cost of change described in the utilitarian theory therefore cannot be explained by increasing returns – on the contrary, it would be more appropriate to speak of decreasing returns here.

55In the second to last section of this article, I have shown that alternative interpretations from historical sociology are not more convincing: [92] factors related to the legitimacy of the instruments, to the power struggles around them or to their functionality cannot in effect account for the endurance of institutions. In the case presented here, the staying power of the instruments can instead be better explained by the role they play in keeping the complexity of the manure management programme in check, for lack of ways to simplify it even partly. Insofar as policy instruments have themselves strongly contributed to increasing complexity, they constitute the keystone of the entire manure management programme; as such, closely studying them is our best chance to understand dynamics of change and/or continuity in public policy. Ultimately, in an effort to complement landmark research by Hall [93] and Jobert [94] on the role of instruments in public policy change, I have emphasized through this case study the role of instruments as an explanatory variable of inertia in public policy.