Simulations of EIA screening across jurisdictions: exposing the case for harmonic criteria?

Abstract

Screening is one of the most critical stages in environmental impact assessment (EIA). However, there is no consensus about best approaches to screen development proposals, particularly in large territories with multiple jurisdictions. This study explored the challenges of harmonizing EIA screening criteria across jurisdictions, using Brazilian states as the empirical context. It set out to answer the following question: what would be the screening decisions of approved development projects if they had been licensed in different states? This study selected four cases of projects that had obtained environmental licenses in the state of Minas Gerais, and, based on their data, simulated license applications in other three states in south-eastern Brazil. Interviews helped to determine likely screening decisions. Screening approaches, while similar, varied significantly in terms of choices of threshold criteria across the four states. With the exception of one project type, the projects would face very different EIA requirements in the other jurisdictions. Such differences seem to stem more from political choices than from geographical idiosyncrasies. Implications for policy-making regarding harmonization are discussed.

Keywords:

• Screening
• environmental impact assessment
• licensing
• developing country
• environmental policy

Introduction

Reports on the state of the environment repeatedly corroborate the world’s growing social and environmental problems (MEA 2005; UNEP 2012; WWF 2014; WWI 2015). It is clear today that government, industry, and civil society need to carefully consider their impacts on ecosystems and society. Environmental impact assessment (EIA), the process of identifying the future consequences of current or proposed actions (Partidário 2012), is one of the world’s key instruments to address this challenge. In 2012, virtually every country on Earth had some sort of mandatory requirement for EIA (Morgan 2012).

The usefulness of EIA as a decision-making tool for sustainability depends on how it is regulated, institutionalized, and implemented. But how exactly to do so remains an open question. Despite growing scholarly interest (Fischer & Noble 2015; Montaño & Souza 2015), one could argue that EIA is still poorly comprehended. Its effectiveness, theoretical foundation, cost-effectiveness, among many other issues, continually ignite debates (Arts et al. 2012; Hanna & Noble 2015). And as Pope et al. (2013) point out, even EIA’s basic concepts, such as significance, screening, scoping, and alternatives, are teemed with unsettled research questions.

Among the most debated issues in EIA is how to determine the potential impact significance of proposed actions and projects (Duinker & Beanlands 1986; Lawrence 2007; Jones & Morrrison-Saunders 2016). Such a challenge is most often perceived in the creation of environmental impact statements (EIS), when developers try to predict their proposal’s most significant impacts. However, as Wood (2008) notes, the determination of impact significance is critical in several other EIA stages, such as screening, scoping, monitoring, and auditing.

Very early in the EIA process, environmental agencies or other competent authorities are required to make a preliminary judgment about the potential impact significance of proposed development actions, in order to decide whether that project should be subject to EIA or other approval requirement (Morgan 1998). This preliminary decision, known as screening, is a key determinant of EIA’s procedural and environmental effectiveness. If potentially ‘(…) harmful activities are screened out, EIA is not able to perform its intended functions’ (Macintosh & Waugh 2014, p. 1). Likewise, if potentially harmless activities are screened in, EIA creates unnecessary burdens, such as ‘(…) additional financial costs and delays for the developer (…) (Wood & Becker 2005, p. 366). The latter scenario is particularly problematic in developing countries.

While there have been significant contributions to the screening literature (Morgan 2012, p. 9), there is still no consensus about best approaches. Context is critical in EIA; the same is true for its screening stage. Nonetheless, the inexistence of one-size-fits-all- approaches should not equate to disregard for common, basic principles. This argument was corroborated by Pope, Bond, Morrison-Saunders and Retief (2013, p. 7) in their 2013 review of the theory and practice of impact assessment. The authors, although acknowledging the complications imposed by the multiple settings under which decisions are made, called researchers to refocus their attention on the basic principles and commonalties that should underpin EIA across different contexts.

This study explored the challenge of developing a more consistent and harmonic approach to EIA screening, using Brazil as the empirical context. More specifically, it set out to answer the following question: what would be the screening decisions of approved projects if they had been subject to EIA in different jurisdictions? To answer this question, this study selected four cases of ‘real’ projects that had obtained environmental licenses in the Brazilian state of Minas Gerais, and, based on their data, created fictional projects to simulate license applications in other three states: Rio de Janeiro, São Paulo, and Espírito Santo. The main purpose of these ‘simulations’ was to understand the rationales behind the screening approaches of each state, in order to identify the key barriers that might emerge in an eventual harmonization of screening procedures and regulations.

The Brazilian EIA system mirrors many of the key traits and challenges found in international contexts. Findings from this study are likely to be relevant to global scholars and practitioners interested in the general topics of EIA screening and policy harmonization. The paper proceeds in six sections. Next, the main approaches to EIA screening are explained. Then the Brazilian context is presented. In Section 4, the methodology is explained. Section 5 presents the results, and Section 6 discusses the implications of the main findings and points out future avenues of research.

Screening approaches

EIA screening has been addressed by peer-reviewed research in different contexts, including Denmark (Nielsen et al. 2005; Christensen & Kornov 2011), India (Rajaram & Das 2011), United Kingdom (Wood & Becker 2005; Weston 2011), Australia (Macintosh & Waugh 2014; Clarke & Menadue 2016), Spain (García-Montero et al. 2010), Netherlands (Enserink 2000) and the whole European Union (EU) (Weston 2004; Pinho et al. 2010). Screening has also been explained in many EIA books (Wathern 1990; Lawrence 2003; Glasson et al. 2005; Sánchez 2008). The literature reveals many approaches to EIA screening. While there is no consensus over taxonomy, the many screening approaches, as illustrated in Figure 1, usually fall into three broad categories: project threshold or policy delineation, case-by-case, and mixed approach.

Figure 1. Approaches to EIA screening. Source: Based on common categories used in the literature (e.g. Canter & Canty 1993; Morgan 1998; Glasson et al. 2005).

The most common screening approach worldwide, according to Glasson et al. (2005), is the threshold approach, also called by Canter and Canty (1993), as policy delineation. Under this approach, decisions to screen projects in or out are based on predetermined threshold criteria delineated in environmental policies and regulations. These thresholds usually mirror criteria such as project type, project size, or environmental/locational factors. This approach often manifests itself in the form of ‘Project Lists’ in the annexes or schedules of regulations. These lists can be qualified as ‘positive’, when they list projects that require EIA; or as ‘negative’, when they list projects that are exempt from EIA. One of the key advantages of the Project List or other threshold approaches is that they are simple and quick to use, and potentially consistent among decision settings. However, these approaches are inherently arbitrary and may not capture contextual information.

In order to overcome the limitations of the threshold or Project List approach, decisions-makers may adopt a case-by-case approach, which is also known as preliminary study or initial environmental assessment approach (Canter & Canty 1993). As shown in Figure 1, this approach encompasses a range of sub-approaches. Government decision-makers may, for example, exercise discretionary judgment about development proposals, taking into consideration numerous factors (Wood & Becker 2005). Decision-makers can also evaluate the proposals’ potential outcomes by engaging with stakeholders (Enserink 2000) or undertaking an analysis of sensitive areas or ecological capacity (Rajaram & Das 2011). While obviously better capable to capture the potential environmental impact significance of development proposals, the case-by-case approach can be complex, slow, and costly.

In addition to the threshold and the case-by-case approaches, decision-makers can adopt a mix of sub-approaches, or even combine case-by-case with threshold screening: a situation that leads to a plurality of options. But, as Lawrence (2003) points out, in any case, effective screening needs to target sustainable development and consider a mix of information derived from proponents, actions and the environment.

Screening acts like a ‘lever point’ in the EIA process, as it can significantly affect the effectiveness of the whole process. Not surprisingly, it is often considered in policy reform. The revised EU EIA directive, for example, introduced relevant changes to screening requirements that, according to Fischer and others (2016, p. 108), are likely to improve screening practices in the United Kingdom. Recent years have witnessed growing debates about the need for change in EIA systems, including its screening methods. This is also the case in Brazil.

EIA screening in Brazil: time for change?

Brazil has a three-decade-old mandatory EIA system underpinned by federal and state legislation. In 1981, EIA and environmental licensing were included among the key mandatory tools of the National Environmental Policy Act. Since then, numerous regulations were issued. Among the most relevant ones are the National Environmental Council (CONAMA) Resolutions 01/1986 and 237/1997, and Complementary Law 140/2011.

The system, however, is being increasingly questioned (Fonseca et al. forthcoming). In the past years, many associations, government and civil society institutions highlighted its problems (e.g. excessive bureaucracy, slow administrative processes, legal insecurity, lack of institutional capacity, among many others), calling for significant changes in EIA regulations (MPU 2004; World Bank 2008; SAE 2009; ABEMA 2013; CNI 2013; FMASE 2013; Hofmann 2015). One could argue that Brazil is on the brink of EIA reform. The National Congress has presented a number of bills that can significantly affect the whole system (Chamber of Deputies 2015; Brazilian Senate 2016). A different, but also impactful proposal has been presented by CONAMA (MMA 2005). EIA screening is one of the areas very likely to be affected.

Development proposals in Brazil can undergo municipal, state or federal-level EIA, but without jurisdictional overlaps. Each Brazilian jurisdiction has its own EIA regulations and procedures. The screening approaches may vary substantially, but they tend to mirror the federal-level approach, which is a clear case of threshold or policy delineation, more specifically of positive Project Lists. The aforementioned CONAMA resolutions 01/86 and 237/97 include lists of projects that specify project types and project sizes subject to EIA and/or environmental licensing. As shown in Figure 2, federal-level screening in Brazil makes a preliminary decision about the potential impact significance of proposals, which can lead to three scenarios: (1) exempt from EIA at the federal level; (2) subject to simplified EIAs; or (3) subject to comprehensive EIAs. Like in many countries, the more likely a project is to have significant environmental impacts, the more comprehensive will be the EIA process. If exempt from EIA at the federal level, proposals can be subject to EIA at lower jurisdictional levels. After Complementary Law 140/2011, there have been growing incentives for municipal-level EIAs. A recent census showed that, in 2015, about 30% of the Brazilian municipalities had some sort of EIA or environmental licensing requirement, especially for small, low-impact projects (IBGE 2016). But, historically and currently, the overall majority of EIA cases fall under state responsibility.

Figure 2. Brazilian federal-level EIA screening. Source: Designed by the authors.

State-level screening is usually more complicated, as state governments can create complementary rules to the federal system. Peer-reviewed studies about EIA screening in Brazilian states are scarce. Numerous questions regarding effectiveness, consistency, evolution, among others, remain largely unexplored. Nonetheless, recent calls for change in screening regulations are corroborating the importance of further investigating this problem. The Brazilian Association of State-level Environmental Agencies (whose acronym in Portuguese is ABEMA) published a report advocating the consideration of locational factors and more objective criteria for significance in screening (ABEMA 2013).

This study is one of the first attempts to understand the challenges of enhancing screening in Brazilian state jurisdictions. It addressed this challenge with a specific focus on harmonization, which has been a preoccupation of large federative countries and political unions (Glasson & Bellanger 2003; Fitzpatrick & Sinclair 2009). This research was partly inspired by the study of Pinho et al. (2010) that critically evaluated the differences in EIA screening across the EU. EU members, although operating under the same EIA Directive, have different screening approaches. One of the study’s main findings was that ‘in general, the harmonization of screening procedures was regarded as advantageous by most of the respondents given the disparities in screening procedures among Member States’ (p. 102). This finding confirmed what the authors had argued: ‘Harmonization contributes to higher levels of cohesion among Member States, facilitating the implementation and enforcement of legislation and regulations at both national and regional levels’.

Brazil has a large territory (more than 8.5 million km²) and a large population (more than 205 million people in 2016). Its 26 states, while operating under the same federal EIA legislation, adopt a variety of screening approaches that create challenges similar to those faced by EU members. Better understanding the degree of consistency in screening across Brazilian states is, therefore, a fundamental step toward the effectiveness of the whole system.

Methodology

In the study of Pinho et al. (2010), the various screening techniques in the EU were investigated through a robust, but somewhat ‘traditional’ methodological approach, based on content analysis, surveys, and interviews. Here, however, the methodology was designed not only to capture regulatory differences across jurisdictions, but also to expose their practical implications to project approval on the ground. This was carried out through a more focused investigation of how different jurisdictions would screen four particular project types (hydropower generation, granite mining, sewage treatment, and gas station) in Brazil’s southeastern region, which encompasses the states of Minas Gerais, Espírito Santo, Rio de Janeiro, and São Paulo. Data collection methods included content analysis, simulation of EIA applications, and interviews.

First, the study undertook a content analysis of the southeastern states’ environmental policies, EIA screening regulations and procedures. Data were obtained in the websites of the states’ environmental agencies, namely: Superintendencia de Regularização Ambiental (SUPRAM), Minas Gerais state’s agency; Instituto Estatual de Meio Ambientel (IEMA), Espirito Santo state’s agency; Instituto Estadual do Ambiente (INEA), Rio de Janeiro state’s agency; and Companhia Ambiental do Estado de São Paulo (CESTESB), São Paulo state’s agency.

To understand the practical implications of the different screening approaches identified in the content analysis, the study undertook simulations of EIA applications. To do so, the study selected four real cases of licensed projects located in the state of Minas Gerais and, based on their actual data, created four ‘fictional’ projects that were very similar to the real ones, except for their locations. These fictional projects were each related to one of those four project types: hydropower generation, granite mining, sewage treatment, and gas station. The study simulated EIA applications for the four fictional projects in the states of Espirito Santo (city of Divino de São Lourenço), Rio de Janeiro (city of Macuco), and São Paulo (city of Borá). The EIA applications for the four fictional projects were undertaken in the selected cities of the other southeastern states, totaling twelve simulations; four simulations in each city (Figure 3). The four real cases of licensed projects located in Minas Gerais were arbitrarily selected because their project types mirror activities commonly found in many Brazilian jurisdictions.

Figure 3. Locations of real and fictional projects in the southeastern region. Source: The Authors.

The simulations of EIA applications were carried out by filling out and submitting environmental license application forms, which were available in the websites of the environmental agencies. In some cases, after submitting the application forms, preliminary screening decisions were communicated to the authors of this study through the states’ online application system. However, to confirm the likely screening decisions, between ago 2013 and December 2014, the authors conducted 15 telephone-based open-ended interviews with customer service representatives, and 13 telephone-based semi-structured interviews with representatives of the environmental agencies of Espírito Santo, Rio de Janeiro, and São Paulo. The codes and profiles of the semi-structured interviews are shown in Table 1.

Table 1. Profiles and codes of the interviewees who helped to determine likely screening decisions.

State	Role in the environmental agency	Environmental agency	Interviewee code
Espirito Santo	Coordinator	IEMA	EES-01
	Coordinator	IEMA	EES-02
	Coordinator	IEMA	EES-03
	Manager	IEMA	EES-04
	Policy assistant	IEMA	EES-05
Rio da Janeiro	Manager	INEA	ERJ-01
	Coordinator	INEA	ERJ-02
	Manager	INEA	ERJ-03
	Environmental analyst		ERJ-04
	Environmental analyst	INEA	ERJ-05
São Paulo	Manager	CETESB	ESP-01
	Manager	CETESB	ESP-02
	Environmental analyst	CETESB	ESP-03

The main objective of the semi-structured interviews, which were anonymous and audio-recorded, was to confirm and/or clarify the main regulations and criteria that would be used in each state to screen the fictional projects in or out, therefore determining likely screening decisions. The interviews helped to understand what licensing and EIA requirements the projects would trigger: licenses, studies, documents, fees, etc. The main findings of the content analysis and simulations were systematically organized in an Excel spreadsheet. Results are presented and discussed below.

Results

Content analysis of screening approaches and significance criteria in state regulations

Findings from the content analysis revealed that the southeastern states’ screening approaches are based mainly on project-based thresholds, whose significance criteria are, in turn, based on projects’ types, size, and biophysical pollution potential. The state of Sao Paulo, for large development projects, also adopts a case-by-case screening approach. Concerns around discretionary judgment, as Wood and Becker (2005) raised in the context of England and Wales, does not seem critical in Brazil’s southeast. The project lists, however, as shown in Table 2, adopt different significance classification codes.

Table 2. Main screening approaches, threshold criteria and project classification codes in the four Brazilian southeastern states.

State	Main screening approaches	Main significance threshold criteria	Significance classes
MG	Positive Project List in the Annex of state regulation DN COPAM 74/04. Rarely case by case approach.	Typology, Size and Biophysical Pollution Potential	Classes 1–6
ES	Positive Project List in the Annex of state regulation IN 12/08, Positive Project List in Annex II of state regulation IN 10/10 and Negative Project List in Annex IV of state regulation IN 12/08. Rarely case by case approach.	Typology, Size and Biophysical Pollution Potential.	Classes S, I–IV
RJ	Positive and Negative Project Lists in Annex 1 of state Decree 44.820/14 and Positive List of the First Article of State Law 1.356/88. Rarely case by case approach.	Typology, Size and Biophysical Pollution Potential	Classes 1–6/A–C
SP	Positive List in Article 57 and Annex 10 of State Law 997/76, Negative Project List on Annex 9 of State Law 997/76, Positive Lists from CONAMA Resolution 01/86 e and State Decree 55.149/09. Eventually, cases-by-case approach for large development projects.	Typology, Size and Biophysical Pollution Potential	Complexity factor (W) and various qualitative descriptions

All four states have regulations that include, within particular articles or Annexes, Positive Lists, that is, lists of projects that should be subject to environmental licensing or EIA at the state level. The states of Espírito Santo and São Paulo, in addition to Positive Lists, use Negative Lists, that is, lists of projects that are exempt from environmental licensing and EIA at the state level. While the policy delineation or threshold approach is the predominant screening approach, all four state governments have discretionary power to use the case-by-case approach whenever necessary. However, the interviewees from the states of Minas Gerais, Espírito Santo, and Rio de Janeiro confirmed that the case-by-case approach is rarely used. In São Paulo, the case-by-case approach is more frequently used for large development projects. Findings from the screening simulations, presented later on, corroborated this situation.

The positive project lists used in Minas Gerais, Espírito Santo, and Rio de Janeiro specify project types subject to EIA as well as threshold criteria used to determine the projects’ pollution potential and size. In these three states, the potential impact significance of the projects are determined through a risk matrix approach, under which the potential impact significance is determined by combining project size and project pollution potential. The higher the size and the pollution potential of the project, the higher the ‘potential impact significance’, which is classified into six different classes. The approaches to determine these six classes vary significantly. For example, Minas Gerais adopts three levels of pollution potential and project size (Table 3), whereas Rio de Janeiro adopts four levels (Table 4). Moreover, the parameters and units used in the projects lists to determine pollution potential and size vary considerably. São Paulo state does not use the risk matrix approach, but a number of different thresholds criteria. One of these is the Complexity Factor, known as ‘W’. The higher the W, the higher the potential impact significance of a project.

Table 3. Classes of potential impact significance used in Minas Gerais’ screening.

		Biophysical pollution potential
		Small	Medium	Large
Project Size	Small	1	1	3
	Medium	2	3	5
	Large	4	5	6

Source: Adapted from COPAM (2004).

Table 4. Classes of potential impact significance used in in Rio de Janeiro’ screening.

Size	Pollution potential
	Insignificant	Low	Medium	High
Minimum	Class 1AInsignificant Impact	Class 2ALow Impact	Class 2BLow Impact	Class 3AMedium Impact
Small	Class 1BInsignificant Impact	Class 2CLow Impact	Class 3BLow Impact	Class 4AMedium Impact
Medium	Class 2DLow Impact	Class 2ELow Impact	Class 4BMedium Impact	Class 5AHigh Impact
Large	Class 2FLow Impact	Class 3CMedium Impact	Class 5BHigh Impact	Class 6AHigh Impact
Exceptional	Class 3DLow Impact	Class 4CMedium Impact	Class 6BHigh Impact	Class 6CHigh Impact

Source: Adapted from CONEMA (2012).

In the four states EIA screening could more accurately be described as ‘Licensing and EIA’ screening, as their objective is to determine whether projects should be subject to environmental licensing, EIA, or both. Potentially low impact projects (e.g. lower Classes or lower W values) may not be subject to EIA, but still trigger the need to submit simpler environmental studies and other licensing requirements. A common misconception about Brazil’s EIA system is that it requires impact assessment studies in every licensing process. Increasingly, however, impact assessment is being waived for potentially low-impact projects. In the lower screening classes of the states of Minas Gerais and Espírito Santo, reviewed here, an environmental license may be granted based on a very simplified licensing process, where developers are required to present a self-declaration of conformity.

In all four states, the social and economic aspects of projects are not considered in regular screening decisions. The ‘pollution potential’ mirror mostly water, air, and soil issues. Social and economic issues can only be considered in case-by-case screening, which rarely occurs. Likewise, the vulnerability of affected areas can only be considered in the unlikely event of case-by-case screening. The Rio de Janeiro and Espirito Santo regulations explicitly require screening decisions to consider a number of locational factors. However, geographic and zoning information systems are still not accurately and reliably integrated into the states’ EIA system, thus limiting the chances of detecting locational constraints.

Overall, findings may suggest the existence of objectivity and consistency in EIA screening across the four states, given the predominance of project-based threshold approaches. However, the parameters and values of project type, size, and pollution potential used in the states’ regulations and procedures are not harmonized, thus leading to different significance thresholds and different licensing, and EIA requirements. The simulations presented below revealed how different those requirements may be.

Simulations of EIA applications for the fictional projects

As described in the methodology section, this study undertook twelve simulations of EIA applications: four simulations in each state. The EIA applications were carried out by filling out and submitting environmental license application forms. Preliminary screening decisions were confirmed through interviews. These ‘experimental simulations’ were particularly challenging in the state of São Paulo, whose EIA regulations, as mentioned above, do not follow the risk matrix approach. In São Paulo, the prediction of EIA screening is more complicated, as discretionary judgment can drive screening to case-by-case analysis. The interviews were helpful to understand likely screening decisions in all states, but especially in São Paulo. Without the inputs of ESP-1 and ESP-3, the likely screening decisions to be taken by São Paulo’s Environmental Agency would be very difficult to predict, particularly with respect to the granite mining and hydroelectric power plant projects. In spite of the difficulties, the authors are confident that the findings, presented in Appendix 1 and discussed below, largely represent the likely screening triggers and decisions that would have taken place if those projects were to be subject to EIA in those three states.

The objective of the simulations were to understand, firstly, what screening criteria would be observed by the environmental agencies and, secondly, what those criteria would trigger in terms of: (a) licensing requirement; (b) type of environmental impact study; (c) need for public hearing; (d) application fees; (e) maximum period for the agencies’ EIA review; (f) license’s maximum renewal period; and (g) need for additional licenses after project is approved. Overall, the findings corroborate the fact that, despite the predominance of threshold approaches within the four states, screening may lead to very different EIA requirements, even when the projects are profoundly similar.

The fictional projects were classified into different ‘potential impact significance’ classes across the states. The parameters and unit of analysis used by state governments to determine project size had various differences. For example, to assess the size of granite mining projects, the agency of Minas Gerais adopted ‘maximum gross production measured in cubic meters per year’, while Espírito Santo’s agency adopted ‘sum of proven reserves measured in hectares’, Rio de Janeiro’s agency adopted several factors including ‘production capacity measured in cubic meters per month’, and São Paulo’s agency adopted ‘mining area measured in hectares and total extraction volume in cubic meters’. Discrepancies like these may lead to different Project Classification and EIA requirements. For example, in Minas Gerais the real granite mining project was classified as a Class 1 project, which represents the lowest impact significance class. In Rio de Janeiro, the project would likely be classified as a Class 3-C (medium impact).

While research has long been showing that screening approaches and criteria are expected to differ among jurisdictions (Pinho et al. 2010), the practical implications of such differences are rarely evaluated. This study helped to address this knowledge gap, by showing that the practical implications to developers can be enormous. In Minas Gerais, the real mining project was subject to a very simple and agile authorization process, which required an application fee of less than US$400.00, no impact assessment study and no public hearing. This project was granted an environmental authorization one day after the application was submitted. The simulation of EIA application in the state of Rio de Janeiro signaled that a similar granite mining project would be classified as Class 3-C project, thus being subject to a far more complicated licensing process. This process would probably require from the developer an application fee of US$28,841.00, the submission of a comprehensive impact assessment study, at least one public hearing, and additional environmental licenses. In Rio de Janeiro, the developer could expect many months, potentially years to obtain the final environmental license required to start production. Such a discrepancy was confirmed by interviewees ERJ-01 and ERJ-02.

The simulations for the Gas Station projects also revealed sharp differences in screening decisions across the states, with profound practical implications. The real Gas Station Project located in Minas Gerais was granted and environmental authorization one day after submitting an application, based solely on a self-declaration of environmental conformity. In the states of Espírito Santo and São Paulo, however, a similar Gas Station Project would probably be required to go through a more demanding licensing process, which requires the submission of simplified environmental studies, longer review periods and additional licenses. The interviewees confirmed such differences.

Appendix 1 also shows numerous discrepancies for the Sewage Treatment Plant. The screenings of hydroelectric projects were an exception, as simulations for this project type revealed a great degree of consistency. The hydroelectric projects, across the four states, would be classified into a higher class of impact significance, thus triggering comprehensive EIA requirements. The main inconsistencies, according to the interviewees, would probably be around application fees, and license renewal periods.

The EIA literature often discusses the limitations of discretionary judgment in screening, arguing that it may lead to inconsistent and unfair treatment of development proposals (e.g. Wood & Becker 2005; Clarke & Menadue 2016). The simulations undertaken here help to show that inconsistency in non-discretionary project lists may also become a problem for businesses that operate in multiple jurisdictions. The profound differences across states in predefined screening thresholds for the Gas Station and Granite Mining project types are clearly creating business incentives and disincentives in particular Brazilian jurisdictions.

The simulations also corroborate the existence, in Brazil, of a recurrent problem found in the literature: the poor consideration of environmental and geographical factors in screening decisions (Rajaram & Das 2011; Slootweb & Kolhoff 2003). The four states analyzed here cannot appropriately consider the vulnerability of affected areas in the screening analysis, as their geographic and zoning information systems are inaccurate and still not integrated into the licensing and EIA system. Screening decisions in southeastern Brazil are still largely based on the project’s features and, therefore, are inherently limited and potentially ineffective. One could argue that ABEMA’s (2013) call for a greater consideration of locational factors in the Brazilian EIA screening system is a reasonable one.

However, ABEMA’s call for greater ‘objectivity’ in the screening system is questionable. The content analysis and the simulations showed that the four states rarely adopt a discretionary case-by-case approach. Screening decisions tend to be based on objective, quantitative criteria, such as project area, flooded area, number of employees, degrees of pollution potential, etc. The states are already adopting objective criteria; however, these criteria may translate into very different and questionable EIA requirements. Therefore, the problem is not so much of lack of objectivity, but rather of lack of ‘accuracy’ in the screening test. The diversity of screening parameters and units used by the four states hints at the lack of technical consensus on how to measure project size and pollution potential, that is, at how to reach a preliminary judgment about the potential impact significance of projects.

In the mid-1980s, Duinker and Beanlands (1986) argued that any exercise in judging the significance of an impact should consider four main factors: (1) the importance of the environment, (2) the distribution of change in time and space, (3) the magnitude of change, (4) the reliability with which changes are measured. The screening approaches of the four states reviewed here tend to address mainly the third factor, and in different ways. Such a narrow and inconsistent consideration of the concept of significance may lead to screening regulations that are vulnerable to criticism. One could even argue that the states’ screening criteria may be reflecting economic and political choices. The fact that the state of Minas Gerais, which houses Brazil`s largest mining reserves, adopts a profoundly streamlined approach to licensing granite mining, may be interpreted as a reflection of this state’s industrial policy. The accuracy of screening criteria and thresholds in Minas Gerais or elsewhere is likely to remain debatable. Without a better and more consensual understanding of the concept of ‘significance’, it is difficult to argue for or against particular regulatory approaches.

Wood and Becker (2005) classified the various screening approaches of competent authorities in the United Kingdom into ‘Precautionary’, ‘Moderate’, and ‘Minimalist’, depending on what percentage of screening decisions led to EIA. The simulations undertaken here shows that the environmental agencies in southeastern Brazil may be precautionary when screening one project type (e.g. hydroelectric projects), but minimalist when screening other (e.g. gas station). It is, therefore, difficult to over generalize the degree of precaution of a particular environmental agency, given that their significance criteria vary across project types.

Final remarks: a case for harmonizing screening criteria?

This is one of the first studies to explore the extent to which screening differences across jurisdictions affect practical business concerns, such as licensing requirements, type of environmental studies, public participation, cost of application fees, and the duration of the process. A somewhat unorthodox methodology, based on simulations of EIA applications was adopted to shed light on these differences. This study corroborates what Joe Weston raised in his review of screening procedures in the European Union, i.e. that the variations in screening practice within jurisdictions that operate under similar regulatory systems ‘(…) stem from the use of different thresholds or triggers for EIA’ (Weston 2011, p. 91). The Brazilian states, while operating under the same requirements of CONAMA resolutions 01/86 and 237/97, adopt complementary regulations with different significance threshold criteria that can translate into very different EIA requirements with enormous practical implication for developers.

The practical implications of screening differences found here expose the existence of ‘tacit’ incentives or disincentives for the installation of particular development projects in some geographical regions. This study showed that granite mining and gas station developers, for example, are more likely to find a faster and less costly EIA process in Minas Gerais than in São Paulo. This situation could provide a competitive advantage for such project types in Minas Gerais. Developers, however, might not be fully aware of such incentives, as they are tacit, ‘collateral effects’ of the states’ screening choices. Futures studies need to confirm if these incentives are actually affecting business decisions in connection with the dozens of different project types subject to EIA in southeastern Brazil.

Such discrepancies of threshold criteria suggest that there is room for regulatory harmonization among Brazilian states. As Pinho et al. (2010) argued, harmonization has the potential to add more cohesion to EIA practice, thus facilitating implementation and enforcement of legislation. In the particular case of southeastern Brazil, harmonization could also mitigate the ‘tacit’ incentives for the installation of specific development typologies.

Legal or regulatory harmonization can bring more consistency to EIA screening across Brazilian states, but consistency does not necessarily translate into accurate screening. As Pope et al. (2013) pointed out, one should not forget the basic principles that underpin EIA. For example, one should not forget that screening is supposed to make a preliminary decision about the potential impact significance of proposals, and, to do so, decision-makers need information not only from projects and actions, but also from likely affected environments. The four states’ environmental agencies reviewed in this study focus their screening attention mostly on the projects’ characteristics rather than on the environment’s vulnerability. Their screening approaches are, therefore, inherently limited. Harmonizing existing screening regulations without ‘fixing’ the states’ geographic and zoning information systems might prove to be unfruitful.

EIA harmonization has been a recurrent concern among European Union members (Lawrence 2003). But, despite many efforts, in 2005, EIA practice still varied widely across Europe (Glasson et al. 2005, p. 312). Even after a recent review of the EIA Directive, EU members still struggled with diverging practices. As a policy objective, harmonization is a difficult one, which requires concerted, long-term efforts.

Brazil has a large territory that houses 26 states and one federal district. Harmonizing EIA screening regulations among states will likely demand strong efforts at federal and state levels. In this context, law and policy-makers need to carefully consider their priorities. Findings from this study suggest that harmonization, while potentially important, does not seem to be as urgent as improving the states’ capacity to consider geographic and locational factors in screening decisions. Future studies should confirm this argument through more focused research about the pros and cons of a number of policy options.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This project was supported by Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) under Grant [473772/2012-4].

Acknowledgments

The authors are grateful to all interviewees and to the two anonymous reviewers for their excellent suggestions and constructive comments.

Appendix 1. Real screening vs. likely screening decisions regarding fictional projects in the states of espirito santo.

Table

HYDROELECTRIC POWER DAM AND PLANT
Environmental Agency	SUPRAM (Minas Gerais)	IEMA (Espirito Santo)	INEA (Rio de Janeiro)	CETESB (São Paulo)
Projects	Real Hydroelectric Project	Fictional Hydroelectric Project 1	Fictional Hydroelectric Project 2	Fictional Hydroelectric Project 3
Affected cities	Conceição do Mato Dentro, Gouveia and Santana de Pirapama	Divino de São Lourenço	Macuco	Borá
Main regulations and respective project lists used in screening	State DN COPAM 74/04	State IN 10/10	State Decree 42.159/09 and INEA Resolutions 52/12 and 53/12	Federal CONAMA Resolutions 01/86 and 237/97
Criteria used to determine likely significance of impacts	Biophysical Pollution Potential and Project Size (based on flooded area and electricity generation capacity in MW)	Biophysical Pollution Potential and Project Size (based on flooded area and length of reduced stream flow below dam)	Biophysical Pollution Potential and Project Size (based on flooded area and electricity generation capacity in MW)	Cases by case analysis based on preliminary environmental assessment by state agency
Significance class and project code	Class 5/Code E-02-01-1	Class IV/Code 21.14	Class 6-A/Code 35.11.02	No specific class number or code. Classified as ‘potentially or effectively capable to cause significant environmental degradation’
Required license	Previous License	Previous License	Previous License	Previous License
Type of environmental impact study	Comprehensive (EIA/RIMA^*)	Comprehensive (EIA/RIMA^*)	Comprehensive (EIA/RIMA^*)	Comprehensive (EIA/RIMA^*)
Subject to mandatory public hearing	Yes	Yes	Yes	Yes
Developer application fee^*	US$ 19,325,90	US$ 9,313.62	US$ 45.356,37	US$ 32,929.00
Maximum period for EIA review	12 months	12 months	12 months	12 months
Enviornmental License Maximum Renewal Period	4 years	5 years	5 years	5 years
Additional Licenses Required if Granted Approval	Instalation License and Operation License	Instalation License and Operation License	Instalation License and Operation License	Instalation License and Operation License
GRANITE MINING
Environmental agency	SUPRAM (Minas Gerais)	IEMA (Espirito Santo)	INEA (Rio de Janeiro)	CETESB (São Paulo)
Projects	Real Mining Project	Fictional Mining Project 1	Fictional Mining Project 2	Fictional Mining Project 3
Affected cities	Itinga	Divino de São Lourenço	Macuco	Borá
Main regulations and respective project lists used in screening	State DN COPAM 74/04	State IN 10/10	State Decree 42.159/09 and INEA Resolutions 52/12 and 53/12	State Law 997/76, Decrees 8.468/76 and 47.397/02, and Cetesb Direction Decision 25/2014/C/I
Criteria used to determine likely significance of impacts	Biophysical Pollution Potential and Project Size (based on maximum gross production measured in cubic meters per year)	Biophysical Pollution Potential and Project Size (based on sum of proven reserves in hectares)	Biophysical Pollution Potential and Project Size (based on serveral factors, such as mining area in hectares, distance from hydric resource in meters, existence of nearby conservation area, need for deforestation, and production capacity, measured in cubic meters per month)	Project Location and Project Size (ased on mining area in hectares and total extraction volume in cubic meters).
Project classification Code	Class 1/Code A-02-06-2	Class II/Code 1.01	Class 3-C (medium impact)/code 00.22.21	No specific class number or code. Classified as ‘very low or non-significant environmental impact’
Required license	Self-declaratory Environmental Authorization (AAF)	Previous License	Previous License	Previous License, if granted authorization by Brazilian Mineral Exploitation Authority (DNPM)
Type of environmental impact study	None. SUPRAM required a self-declaration of conformity.	Simplified EIS (RAP)	Comprehensive EIA (EIA/RIMA)	Simplified EIS
Subject to mandatory public hearing	No	No	Yes	No
Developer application fee^*	US$ 368.11	US$ 161.98	US$ 28,841.50	US$ 3,675.97
Maximum period for EIA review	3 months (but took only 1 day)	6 months	12 months	6 months
Enviornmental license maximum renewal period	4 years	5 years	5 years	2 years
Additional licenses required if granted approval	None.	Previous License and Operation License	Previous License and Operation License	Previous License and Operation License
SEWAGE TREATMENT PLANT
Environmental agency	SUPRAM (Minas Gerais)	IEMA (Espirito Santo)	INEA (Rio de Janeiro)	CETESB (São Paulo)
Projects	Real Sewage Treatment Plant Project	Fictional Sewage Treatment Plant Project 1	Fictional Sewage Treatment Plant Project 2	Fictional Sewage Treatment Plant Project 3
Affected cities	Caratinga	Divino de São Lourenço	Macuco	Borá
Main regulations and respective project lists used in screening	State DN COPAM 74/04	State IN 10/10	State Decree 42.159/09 and INEA Resolutions 52/12 and 53/12	State Law 997/76, Decrees 8.468/76 and 47.397/02, and Cetesb Direction Decision 25/2014/C/I
Criteria used to determine likely significance of impacts	Biophysical Pollution Potential and Project Size (based on average flow measured in Liters per second)	Biophysical Pollution Potential and Project Size (based on project maxium flow measured in Liters per second)	Biophysical Pollution Potential and Project Size (based on extension of sewage lines in quilometers, quantity of sewage pumping stattions and treatment flow in liters per second).	Project size (based on total attended population)
Significance class and project code	Class 3/Code E-03-06-9	Class III/Code 19.02	Class 2-B (low impact)/Code 35.41.14	No specific class number or code. Classified as ‘very low or non-significant environmental impact’
Required license	Concomitant Previous License and Installation License	Previous License	Simplified Environmental License	Concomitant Previous License and Installation License
Type of environmental impact study	Simplified EIS (RCA and PCA)	No impact assessment study, but required various documents, projects and environmental baseline information)	No impact assessment study, but would be required to submit various documents and projects)	No impact assessment study, but would be required to submit various documents and projects)
Subject to mandatory public hearing	No	No	No	No
Developer application fee^*	US$ 5,339.84	US$ 512.92	US$ 863.17	US$ 27,345.61
Maximum period for EIA review	6 months (but took 80 days)	6 months	6 months	6 months
Enviornmental license maximum renewal period	6 years	5 years	10 years	3 years
Additional licenses required if granted approval	Operation License	Instalation License and Operation License	None	Operation License
GAS STATION
Environmental agency	SUPRAM (Minas Gerais)	IEMA (Espirito Santo)	INEA (Rio de Janeiro)	CETESB (São Paulo)
Projects	Real Gas Station Project	Fictional Gas Station Project	Fictional Gas Station Project	Fictional Gas Station Project
Affected cities	Ribeirão das Neves	Divino de São Lourenço	Macuco	Borá
Main regulations and respective project lists used in screening	State DN COPAM 74/04	State IN 10/10	State Decree 42.159/09 and INEA Resolutions 52/12 and 53/12	State Law 997/76, Decrees 8.468/76 and 47.397/02
Criteria used to determine likely significance of impacts	Biophysical Pollution Potential and Project Size (based on fuel storage capacity measured in cubic meters)	Biophysical Pollution Potential and Project Size (based on fuel storage capacity measured in cubic meters)	Biophysical Pollution Potential and Project Size (based on production and storage area measured in square meters, and number of employees)	Project type
Significance class and project code	Class 1/Code F-06-01-7	Class III/Code 27.01	Class 2-C (low impact)/Code 55.21.31	No specific class number or code. Classified as ‘very low or non-significant environmental impact’
Required license	Self-declaratory Environmental Authorization (AAF)	Previous License	Simplified Environmental License	Previous License
Type of environmental impact study	None. SUPRAM required a self-declaration of conformity.	Simplified Environmental Impact Study.	Simplified Environmental Impact Study.	No impact assessment study, but would be required to submit various documents and projects)
Subject to mandatory public hearing	No	No	No	No
Developer application fee^*	US$ 368.11	US$ 512.92	US$ 1,094.2	US$ 249.23
Maximum period for EIA review	3 months (but took only 1 day)	6 months	6 months	6 months
Enviornmental license maximum renewal period	4 years	5 years	10 years	2 years
Additional licenses required if granted approval	None.	Previous License and Operation License	None.	Previous License and Operation License

^*To facilitate interpretation of application fees in an international context, the fees in Brazilian Reais were converted to American dollars, adopting a conversion rate of 3.