Sustainable Rust Belt Manufacturing: An Exploration of Emission Rates during the Enforcement Era of the Great Lakes Water Agreement Act (GLWQA)

: The Great Lakes as a key global resource provides an abundance of freshwater as well as power for the entire region, but environmental degradation due to industrial discharge into the water supply has depleted the Great Lakes ecosystem. The Great Lakes Water Agreement Act (GLWQA) is a framework to protect and restore this ecosystem. It has changed in scope since its inception, and while the regulatory duties of the GLWQA have been noble, less input from Rust Belt industry and more power from the steering committee of the GLWQA have resulted in power struggles among its various stakeholders as deindustrialization has spread across the region. This study will analyze different eras’ amendments to ascertain changes in water pollution as they relate to production output rates over the past generation in order to analyze the effects of the various rounds of water pollution enforcement versus Rust Belt industrial output compared to other states in the country.


INTRODUCTION
The Great Lakes, consisting of Lake Superior, Lake Michigan, Lake Huron, Lake Erie and Lake Ontario, are the largest group of freshwater lakes on earth and contain 21% of the earth's freshwater.Of the 40 million gallons used daily from the Great Lakes, more than half is used for industrial power production (Environmental Law and Policy Center, 2019), and 95% of the freshwater used as drinking water in the US comes from the Great Lakes (The 71percent, 2020).The Great Lakes has been called "unquestionably a unique national and global resource" (Freedman & Monson, 1989, p. 285).Truly, the Great Lakes ecosystem is a vitally important aspect of American functionality.The Great Lakes is today a region by itself constituting the world's third-largest economy (Desjardines, 2017).Martin (1999) stated that the Great Lakes includes "one of the world's largest concentrations of industrial" output (p.15) and the Environmental Law and Policy Center (2019) noted that "agriculture, industrial manufacturing, fishing, and recreation together form an economic engine" (p. 1).The Great Lakes has contributed heavily to the economic prowess of the Rust Belt geographic region.This area, also referred to as the Manufacturing Belt, consists of Midwest American states, most of which border the Great Lakes.The Rust Belt's rise in economic strength during the late nineteenth and early twentieth centuries is attributed to the coinciding rise of the manufacturing sector (Tanoos, 2010;Stiglitz, 2017).Many believe that this region, powered by the coal industry, helped the North win the Civil War and propelled the United States into global hegemony in the industrial era (Cooke, 2006;Biggers, 2014).Unfortunately, during the past several decades, the decline of US manufacturing has been accompanied by job loss attributed to plant closings in this region, providing the impetus for its more recent nickname "Rust Belt" (Deakin & Edwards, 1993;Chase, 2003;Brown, et al., 2008;Tanoos, 2010;Bernero & Peduto, 2016).Reksulak et al. (2013) found that Rust Belt states witness comparatively higher industrial costs due to regulations compared to other areas of the country.Organizations in the Rust Belt have often blamed these regulations for inhibiting the region's industrial effectiveness (Isenberg, 2017).In particular, as Cooke (2006) and Biggers (2014) noted, since the Rust Belt is generally powered by the coal industry, fossil fuel regulations have especially negatively affected these economies.As such, Rust Belt stakeholders have been generally opposed to direct and indirect federal regulations.Although deindustrialization-related job losses have been a negative trend associated with the Rust Belt, there is a potentially more pressing issue, being the coinciding phenomenon of the degradation of the ecosystem's vital water supply (Ahamad et al., 2021).Fischer (2003, p. 51) said that the Great Lakes especially has "been acutely vulnerable" to environmental degradation over the years.Water pollution constitutes discharged pollution into natural water bodies (Mambretti and Jimenez, 2020); past industrial activity has been found to be the largest culprit in the degradation of the quality of the Great Lakes' freshwater.In addition, climate change is negatively affecting the Great Lakes region more than other regions (Spring, 2001;Egan, 2017;Dempsey, 2019; Environmental Law and Policy Center, 2019; Crossman et al., 2020).Manufacturing in the Great Lakes has led to the industrial discharge into the water supply of chloride sources and other pollutants such as garbage disposal waste, sewer runoff and water softening products, along with other pollutant sources from chemical, steel, and food manufacturers (Sonzogni et al., 1983).As a result of past decades of industrial discharge into the Great lakes, the Great Lakes ecosystem has massive biological "dead zones" (Egan, 2017) where ecology is not active.Today, toxic substances in the Great Lakes aren't usually directly discharged directly into it but instead originate from regional landfill leaks and/or contaminated sites that eventually find their way to the Great Lakes (Carpenter, 2007).The degradation of the quality of the freshwater in the Rust Belt and Great Lakes region has resulted in governmental regulations to limit water pollution.The US and Canada jointly passed the binational Great Lakes Water Agreement Act (GLWQA), enacted in 1972.Swain (1981) surmised "Both the United States and Canada have long recognized the importance of their boundary waters, and the need to preserve them as a priceless international resource and heritage" (p.447).When it was first passed, Fischer (2003, p. 52) noted that regulatory means were part of water pollution control because past reductions in water emissions into the Great Lakes ecosystem have been due to "both to voluntary initiatives and regulatory controls on industrial sources" (Fischer, 2003, p. 52).The GLWQA began as a way to implement goals and actions for improving water quality in the Great Lakes and has been amended several times.It has been called the "centerpiece of the institutional architecture to address water quality in the Great Lakes" (Berardo et al., 2019, p. 9).Private industry has been part of the guidance and writing process throughout, particularly in the leadup to the GLWQA (Martin, 1999; Grover and Krantzberg, 2018).While industries have been involved in the past development of the agreement (Freedman & Monson, 1989), an amendment, The Great Lakes Water Quality Initiative of 1990, involved more federal regulatory agencies including the EPA and less input from Rust Belt-based organizations, especially in the steering committee (National Water Quality Inventory, 1994).Industry eventually labeled it as too "stringent" because of its "requiring the states to adopt strict standards for waste disposal and discharge into the lakes" (Cox, 2013, p. 122).In 2001, the Great Lakes Charter Annex, a handshake agreement between governors of Rust Belt states and premiers of Canadian provinces, was formed to better coordinate the management of the Great Lakes water supply.The Annex received even more ardent opposition from Rust Belt industry (Inside Washington Publishers, 2004).In 2008, the Great Lakes Compact was signed between eight US states bordering the Great Lakes in order to better manage the water supply (Annin, 2018), with minimal input from private industry.Current solutions involve the lengthy and expensive process of dredging of shipping channels to alleviate issues such as massive amounts of toxic algae (Egan, 2017).As of 2022, the US Environmental Protection Agency (EPA) coordinates the American responsibilities within the GLWQA.In spite of opposition, much has been done over the past decades to mitigate water pollution in the Great Lakes.Some critics say that regulations enacted on manufacturers in the Great Lakes need to go further to alleviate water degradation.Other pundits point out that the negative effects of climate change will hurt Great Lakes manufacturers more in the future than increased regulations in the present (Shin, 2013).However, contrary viewpoints say that regulations have gone too far and have inhibited efficient models of production (Beecher & Kalmbach, 2013; Keiser & Shapiro, 2019).A 2017 report noted that "the International Joint Commission, a US-Canadian panel that monitors Great Lakes water quality, states that efforts to clean up the lakes over 'the past 25 years are 'a mix of achievements and challenges'" (McCartney, 2017).By 2019, public and private US sources had spent nearly $5 trillion in efforts to keep water in the US clean.In addition, Keiser and Shapiro (2019) noted that people may be willing to pay more in taxes to help keep clean these "iconic waters" (p.66).Mambretti and Jimenez (2020) surmised that reaching solutions to global water pollution requires collaboration and an overall interdisciplinary approach, and experts have noted that the GLWQA certainly has room to evolve in its scope (Devy and Davis, 2020; Hale and Anderson, 2021).To ascertain how different amendments to the GLWQA have altered water pollution and production output rates over the past generation, this study will analyze data from Rust Belt states that discharge pollutants into the Great Lakes.The effects of the various rounds of water pollution enforcement versus industrial output will then be compared to those of other states in the country.

METHODOLOGY/RESULTS
The data for this study were mined using a cross-sectional analysis to assess water pollution resulting from industrial productivity/output in conjunction with several key GLWQA milestones.Therefore, this study focused on datasets from the first and subsequent years in which various water pollution regulations were enforced, including 2002 (one year after the GLCA was enacted), 2014 (the first year in which the GLWQA was fully enforced and implemented), and 2018 (one year after the International Joint Commission progress report and edicts).
The sample set of Rust Belt states was determined as follows: states that border the Great Lakes but are not generally considered to be part of the Rust Belt (New York to the east and Minnesota to the west), and those with comparatively smaller Great Lakes coastlines (Indiana with 45 miles and Illinois with 63 miles) were omitted, as was Iowa, a Rust Belt state with no coastline.That left Ohio (312 miles), Wisconsin (820 miles), Michigan (3,224 miles), and Pennsylvania (140 miles) as the sample set of Rust Belt states, with particular focus on Michigan due to its extensive coastline (Office for Coastal Management, 2020).
The data for total on-site and off-site disposal or release of chemicals (total pollution rates) were mined from the US Environmental Protection Agency's (EPA) Toxic Release Inventory (TRI) from the most recent year (2021) (EnviroEPA, 2022).The TRI is a publicly-available EPA database containing information on the release of toxic chemicals (Antisdel, 2017).In order to assess economic data specific to industrial output, statistics from the US Bureau of Economic Analysis (BEA) were used to ascertain annual state GDP specifically related to the manufacturing process, or "Annual Gross Domestic GDP by state, GDP current dollars" (Bureau of Economic Analysis, 2021).Then, a cross-sectional analysis of the data from these datasets was completed to create two indexes.
First, to obtain a comparable method for assessing pollution related to manufacturing output, GNP was divided by water pollution to produce a pollution efficiency index (PEI); a larger PEI would indicate a more sustainable manufacturing process in terms of pollution (Tanoos, 2021).A PEI can determine successful states that manufacture at high rates while also polluting at low rates.A closer examination finds extreme values in the first two sample sets whose PEI was larger than 2,000,000 in at least one year: Arizona, Massachusetts, New Hampshire, and Rhode Island.These were deemed to be outliers.1).For this index, since the p-values were larger than 0.05, it can be claimed that the distributions of the two groups are roughly equal, and thus there is no significant difference between the two groups.

𝑷𝑬𝑰 = 𝑮𝑵𝑷 𝒕𝒉𝒆 𝒂𝒎𝒐𝒖𝒏𝒕 𝒐𝒇 𝒘𝒂𝒕𝒆𝒓 𝒑𝒐𝒑𝒖𝒍𝒕𝒊𝒐𝒏
The results of the f-tests showed that the variances of the Rust Belt versus non-Rust Belt sample sets were different; specifically, the p-value for the f-test was less than 0.05, the level of significance (see appendix A).This result suggests that the Rust Belt sample set produces more efficiently in terms of water pollution.
Next, to find an index to obtain a comparable method for assessing water pollution related to total pollution through the different eras of the GLWA's water pollution regulations, variables for both water pollution and total pollution were utilized.This metric of water pollution divided by total pollution, or "PCT", takes into account other types of pollution such as air pollution, soil  The results of the full sample set from Table 3 suggest that there is no significant difference between the two full sample sets during these years.However, when extreme values were omitted, the PCT p-values in the Welch test were less than 0.05, indicating that there is a statistically significant difference between the sample sets (see Table 4).As such, it can be claimed that the Rust Belt states witnessed increased water pollution in 2014 compared to the rest of the country.

REACTIONS/FUTURE STUDIES
Rust Belt states were polluting the water at higher rates than the rest of the country, but not if manufacturing-related GNP was included.While the Rust Belt did pollute their water supplies at increasing rates compared to the rest of the country, that difference has tapered off.In 2014, over 6.92% of the pollution coming out of the Rust Belt came from water pollution, compared to 3.55%

Figure 1 .
Figure 1.Pollution Efficiency Index equation pollution, and so on, so that the PCT is representative of changes in water pollution versus these other forms of pollution.A lower PCT is more ideal because it represents comparably low water pollution and/or decreasing water pollution.

Table 1 .
Average PEIs of Rust Belt versus Non-When omitting those four outlier states in the non-Rust Belt sample set, the Rust Belt states were producing comparably more efficiently in terms of pollution in 2002 (33,060.3PEIversus14,941.8PEI)butby 2018, they were producing 3.8 times less efficiently (19,970.9versus75,349.9)comparedto the non-Rust Belt states.By 2018, the Rust Belt's production efficiency was similar to that in 2014(27,564.6versus74,673.7,or 3.71 times more efficient) (see Table

Table 2
below shows the PCT for both sample sets (the four extremes were removed).

Table 2 .
Average PCTs of Non-Rust Belt/Rust Belt (no extremes)

Table 2 ,
the Rust Belt states witnessed a sharp increase in PCT in 2014, which then leveled off in 2018.In contrast, the non-Rust Belt states saw a PCT which decreased over time.The f-test was utilized to verify that the variance of two groups (state1=1 and state1=2) were unequal, resulting in less than .05for2014and 2018.If the variances were the same, a standard t-test would have been utilized.Since the sample sets had unequal variance, or spreads of all the data points for each set, the next step was to test for significance of the disparity between the sets.deWinter(2013) suggested that the Welch test provides advantages over other tests of unequal variances, especially with unequal sample sizes, and when one sample was drawn from a small population as was the case with the Rust Belt sample.As such, this study applied both the Welch test and the Mann-Whitney U test, as shown in Table3.

Table 4 .
PCT: P-values when extremes removed p-value of different tests when comparing PCT (Rust Belt versus Non-Rust Belt)