Do voluntary pollution reduction programs help reduce pollution levels? Evidence from the Mexican Clean Industry Program

This paper evaluates the effectiveness of the Mexican Clean Industry Program, in which firms participate voluntarily if they are willing to meet the pollution emissions legal standards in exchange for a Clean Industry Certificate. It first develops a simple model with two groups of players, firms and the authorities. Firms can choose to be in compliance, non compliance or participate in the program. Authorities can set the cost of non compliance by changing the frequency with which they inspect different industrial sectors. By imposing some structure to the cost of participation and the cost of compliance, we show evidence from aggregate data at the industrial sector level suggesting that firms with relatively low cost of compliance are the ones that participate in the Mexican Clean Industry Program. However, as authorities have the option to update the inspection intensity given the number of firms participating in the program, certification serves as a screening tool that reduces the cost of inspection in sectors with a high percentage of certified firms. According to our model, the reductions in pollution emissions levels are not only observed amongst participating firms, but also amongst non-certified firms in industrial sectors with a high percentage of certified firms. The predictions of the model are further tested with firm level data. For each firm’s exact geographic location (zip code), a monthly measure of particulate matter in the atmosphere obtained from satellite imagery is assigned for the period between March, 2000 and December, 2006. While particulate matter concentrations seem to lower significantly where certified firms are located, a significant reduction is also observed in places where non-certified firms in sectors with a high percentage of certified firms are. This last relationship is not observed for firms with less than 10 employees, which are not subject to inspections by the authorities. JEL codes: Q52, Q56. Introduction. The evaluation of the pollution control policies in developing countries, especially those with high levels of trade with industrialized nations, is becoming increasingly necessary. The fear of the “pollution haven” hypothesis, on one hand, which claims that weak regulation in developing countries represents an incentive to high polluting industries to relocate in poorer regions when the barriers to trade are low and, on the other hand, the possibility of trade creating a “race to the bottom”, in which competition between nations to attract or retain firms incentives them to relax their pollution control policies, suggest a high need to precisely assess countries’ success at controlling pollution emissions. Given the high levels of trade between the US and Mexico, the potential differences in the effectiveness of law enforcement between the two countries and the decreasing trade barriers between them, a careful evaluation of the effectiveness of the Mexican authorities at controlling pollution emissions seems then of special relevance for the US, Mexican and international contexts. This paper contributes to the literature evaluating the policies that the Mexican authorities have put in place in order to control pollution emissions by firms, focusing on the two main policy tools used by the Mexican Federal Environmental Protection Agency (Procuraduría Federal de Protección al Ambiente, PROFEPA). On one hand, this agency is responsible for inspecting firms in order to determine if they comply with the current legal pollution emission standards. Inspections are performed at random, assigning a higher probability of inspection to sectors with higher perceived risk of polluting. If a firm is found to be in non compliance, it is forced to pay a fine, which increases in case of relapse. Relatively small firms seem to be rarely subject to inspections. On the other hand, the same agency, in 1997, introduced the Mexican Clean Industry Program (Programa de Industria Limpia), also known as National Environmental Auditing Program (Programa Nacional de Auditoría Ambiental), the main voluntary pollution reduction program in Mexico. Firms participating in this program have to pay for an audit by an independent agency that determines the actions to be taken in order to comply with the pollution emissions standards and, after they succeed at meeting the pollution levels standards, they are granted a Clean Industry Certificate, which can be used for marketing purposes. If certified, firms are exempted from inspections for a given period of time (at least two years). Since 1997 until 2007, 2,568 firms had received this certification. This paper argues that these two policies cannot be evaluated in isolation. In general, it makes two main contributions. On one hand, it suggests a simple framework, which models both the authorities’ inspection policy and the firm’s decision between participating in the Clean Industry Program, being in compliance with pollution emission standards and being in non compliance and subject to fines, that allows the researcher to test for the characteristics of firms that have been granted a Clean Industry Certificate without the need of detailed firm level data (which seems to be one of the limitations for empirical studies). On the other hand, it stresses the fact that, even if relatively clean firms are the ones participating in the program, its effectiveness, when only looking at participating firms, might be understated. If used together with an inspection policy, the introduction of voluntary programs can reduce the cost of inspections by revealing information to the authorities about participating and non participating firms’ characteristics. Authorities (can) set a higher inspection rate for non participating firms at a lower cost when a high percentage of firms get certified, increasing the incentives for non participating firms to reduce their pollution levels. This last fact seems widely ignored in the empirical literature. The paper is presented as follows. The next section briefly reviews the empirical literature trying to evaluate the effectiveness of programs of this kind, and motivates the need for a theoretical framework for its evaluation. Section III describes the simple setup that we develop in order to describe participation in the program and the authority’s role in terms of its inspection policy. Section IV describes the aggregate data and discusses its implications in terms of the characteristics of participating firms. Section V returns to the model and explores how the authorities can use the information revealed by participation in the program to update the inspection policy and tests the model’s predictions with the aggregate data. Section VI describes the firm level data set used to further test the predictions of our empirical model, and the empirical strategy. Section VII shows the empirical results. The last section concludes. II. Motivation. Voluntary pollution reduction programs, similar to the Mexican, are popular tool used by policymakers around the world as part of the instruments aiming to encourage firms to reduce their emissions levels (OECD, 1999, 2003). Their emergence has been followed by a growing body of literature trying to evaluate their effectiveness, generally studying these programs in isolation from other environmental protection actions. Given this, we believe that most of the empirical studies trying to evaluate voluntary pollution reduction programs fail at taking into consideration a more general equilibrium perspective on their effectiveness. By looking at these programs in isolation, they ignore the relevance of the information revealed in the process of certification, and how that information can be used by other actors and influence firm’s behavior. The existing literature seems especially concerned with testing if participating firms are those already in compliance with the emissions standards, or if firms invest in pollution reduction for reasons not related to the existence of the program (Vidovic and Khanna, 2007; Morgenstern and Pizer, 2007) and it has focused on industrialized countries. Specifically, the US Environmental Protection Agency 33/50 program has received most of the attention in the literature. Arora and Carson (1996), GamperRabindran (2006) and Sam and Innes (2006), for example, do not find evidence that firms participating in the program were those who had reduced their emissions before the implementation of the Program. However, Vidovic and Khanna (2007) find the opposite. According to them, a very small percentage of the total emissions by participating firms can be attributed to the program. Some of these studies also try to test if the firms with the lowest or highest emissions levels are the ones participating, with no conclusive results. The differences in the findings seem to come from differences in the sample used, the possibility for correction for selection into the participating sample, or the variable used to measure environmental compliance (Alberini and Segerson, 2002). There exists one study trying to evaluate the Mexican Clean Industry Program (Blackman et al., 2007). It shows that firms that have been inspected or fined for not complying with pollution emissions standards in the past are more likely to participate. It argues that this is evidence that the program is contributing to reduce pollution, given that participating firms are more likely than average to be in non compliance before entering the program, and presumably in compliance with environmental regulations when they graduate from it. However, as we will see later, the positive correlation between the probability of being inspected by the authorities and participation in the program might prove that it is the relatively cleaner firms that decide to participate in the Clean Industry Program, contrary to what is argued by the mentioned study. Both the empirical challenges faced when trying to determine if it is the cleanest firms that are getting certified, and the possibility for other actors’ responses to participation, which can incentive non participating firms to change their behavior as a consequence of the existence of programs of this kind, motivate the need for a theoretical framework, which we develop in the following section. The model will allow us first to test if the cleanest firms are the ones participating in the program from aggregate data and it will help us argue that any paper trying to evaluate the overall impact of voluntary pollution reduction programs should not only focus on participating firms, but also on the authorities’ and non participating firms’ response to participation. We will show that it is possible to observe no reduction in pollution emissions for participating firms, but rather an increase in the incentives for non participating firms to improve their environmental performance. III. Modeling firms’ participation in the Clean Industry Program. III.1. The Firms’ Problem. A first glance at aggregate data at the sector level for inspections performed by PROFEPA, inspections resulting in non compliance and participation in the Clean Industry Program seems rather puzzling. As can be seen in Table 1, the correlation between the percentage of firms inspected in each industrial sector and the non compliance rate (for the whole 1992 to 2006 period) is very close to zero, while the percentage of firms in each sector that have received a Clean Industry Certificate and the same probability of inspection are strongly positively correlated. Our model will explain these relationships, arguing that the lack of correlation between inspections and non compliance can be evidence of the effectiveness of the Mexican inspection policy, which imposes higher incentives to potentially polluting firms to be in compliance, as they are assigned a higher inspection probability. On the other hand, as we will see, when imposing some structure to the cost of compliance and participation in the Program, the positive correlation between inspection intensity and certification rate can provide information about the types of firms within industrial sectors that get certified.