Assessing the flow to low-income urban areas of conservation and environmental funds approved by California’s Proposition 84

Assessing the flow to low-income urban areas of conservation and environmental funds approved by... a1111111111 a1111111111 Government funding accounts for a large proportion of conservation and environmental improvements, and is often the result of citizen votes on state ballot measures. A key con- cern surrounding public investments in the environment is whether that funding serves lower-income communities, which are often the communities of greatest need. We applied OPENACCESS three statistical methods to analyze the spatial distribution of conservation funding derived from California’s Proposition 84, which distributed nearly $4 billion across California Citation: Davies IP, Christensen J, Kareiva P (2019) Assessing the flow to low-income urban between 2006 and 2015. First, we used hurdle models to ask if income, population density, areas of conservation and environmental funds urban coverage, or pollution could explain receipt of grants or magnitude of funding. Sec- approved by California’s Proposition 84. PLoS ONE ond, we compared the income levels of funded and unfunded communities for each chapter 14(2): e0211925. https://doi.org/10.1371/journal. of the proposition. Finally, we examine two sections of the proposition that were intended to pone.0211925 fund parks around the state and compare the attributes of funded and unfunded communi- Editor: Jacint Balaguer, Universitat Jaume I, SPAIN ties. Proposition 84 offers lessons for environmental legislation and future research. While Received: March 15, 2017 there were general tendencies for more funding to flow to poor areas and areas with pollu- Accepted: January 24, 2019 tion problems, the language in Proposition 84 as a whole was vague with respect to the Published: February 7, 2019 funding of disadvantaged areas, and as a result the targeting of these areas overall was at best modest. However, when enabling legislation (AB 31) defined specific “metrics of disad- Copyright:© 2019 Davies et al. This is an open access article distributed under the terms of the vantage” that had to be met by communities to receive funds from some sections of Proposi- Creative Commons Attribution License, which tion 84, the funds did flow much more selectively to poorer communities. This suggests that permits unrestricted use, distribution, and future ballot measures should be very explicit in their language if they want to promote equity reproduction in any medium, provided the original author and source are credited. in conservation investments, and that future research should investigate the extent to which technical workshops and outreach could further increase the number of funded grant pro- Data Availability Statement: All relevant data are within the paper and its Supporting Information posals from low-income communities. files. Funding: Jon Christensen received funding from Resources Legacy Fund, #11026 (http://www. resourceslegacyfund.org/). Ian P. Davies and Peter Kareiva received funding from the Anthony and Introduction Jeanne Pritzker Family Foundation. In the United States, ballot measures have become one of the largest funding mechanisms for Competing interests: The authors have declared that no competing interests exist. public investment in conservation and environmental improvements. In the last decade, $40 PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 1 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure billion of funding has been approved through ballot measures for conservation alone [1]. When a ballot measure or funding program is approved, headlines typically announce a vic- tory for the environment and a victory for conservation [2]. However, critics of these ballot measures have pointed out that the funding often fails to address social inequities in environ- mental protection and access to parks and natural areas [3,4]. Although minority groups such as Hispanics and African Americans consistently vote in favor of environmental measures, if these measures are seen to reinforce social inequality, this key support for conservation and environmental protection may be lost [5]. Thus it is increasingly important to examine conser- vation and environmental funding from an environmental justice perspective and ask how well low-income and high-need communities are served [6,7]. In their analysis of the 1996 Proposition K in Los Angeles, Wolch et al. found that park funding from this public ballot measure often compounded existing inequalities in park access by funding park improvements rather than investments in new properties [3]. Other studies suggest that funding often lacks a focus on structural inequity in park access; that is, urban residents tend to use park space more intensively than their rural and suburban counterparts, a fact that is ignored when measuring access by park space per unit area rather than per active user [3,8,9]. Following the money is important because if there are particular social or environmental needs that are not receiving funds, adjustments could be made for future policy. Yet to our knowledge, such a quantitative analysis has never been performed for a state ballot measure. Throughout California’s history, a number of ballot measures have been passed with the intention to fund environmental projects. These range from local measures like Proposition K in Los Angeles to the statewide Propositions 12 and 13 which all, in varying ways, intended to channel public funds to increasing park space and improving access. Minorities tend to live in cities with less local fiscal capacity to spend on parks, and for cities in California, public funds from state legislation and ballot measures have become a viable model for building infrastruc- ture [10]. Large environmental nonprofits play an important role in this process by helping to craft these measures through political partnerships and then donating to them so that they are passed [11]. While successful in passing park measures, some have expressed concern that nonprofits whose concern is habitat protection will prioritize green spaces on the edges of cit- ies, rather than in the urban core where few have access to open spaces [10]. When these deci- sions are made in writing measures and awarding grants, they can lead to a distribution of park resources that is not equitable for the communities most in need. An Overview of Proposition 84 Proposition 84, a general obligation bond, was passed in 2006 and at the time represented the largest state ballot measure in the United States for environmental protection. Notably, it was carried largely with support from California Latinos, who voted 84% in favor versus just 45% from non-Hispanic white voters [12]. Proposition 84 authorized $5.4 billion in spending on water quality and supply, natural resource protection, and urban greening in high-need areas– a wide breadth, leading some to criticize the measure for lacking clarity and accountability [13,14]. Others criticized the explicit earmarking of funds to specific groups, like the San Joa- quin River Conservancy, as evidence of too much sway from environmental donors [15]. Proposition 84 funded projects through a competitive grant process. Under “general provi- sions” the text of Proposition 84 specified the following social priorities: “assistance to commu- nities with contaminated sources of drinking water” and “revitalizing our communities and making them more sustainable by investing in . . . local parks and urban greening” [16]. The $5.4 billion was then divided by project type into nine distinct chapters, each with its own grant criteria for funding (Table 1). PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 2 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Table 1. Proposition 84 fund allocation from 2006–2015. Grants with local impacts are those projects we determined to have identifiable “on-the-ground” impacts in local communities, as opposed to planning and technology grants and large regional projects where local community impacts could not be identified. Chapter 1 of the bal- lot measure details general provisions of the proposition but has no funding tied to it, so it is excluded from this analysis. Chapters Funding Authorized Funding Awarded (in millions of dollars and number of grants) (millions) Grants with local Grants for which impacts are not readily assigned to All impact particular locations grants 2: Safe Drinking Water and Water Quality $1,525 $143 $1,003 $1146 Projects 173 796 969 3: Flood Control $800 $182 $411 $594 36 342 378 4: Statewide Water Planning and Design $65 $0 $64 $63 0 17 17 5: Protection of Rivers, Lakes, and Streams $928 $524 $132 $656 594 399 993 6: Forest and Wildlife Conservation $450 $299 $26 $325 221 30 251 7: Protection of Beaches, Bays, and Coastal $540 $161 $140 $301 Waters 249 266 515 8: Parks and Nature Education Facilities $500 $236 $67 $303 571 118 689 9: Sustainable Communities and Climate $580 $419 $102 $521 Change Reduction 308 256 564 Total $5,388 $2,152 $1,946 $3,909 2152 2224 4376 https://doi.org/10.1371/journal.pone.0211925.t001 While Proposition 84 was intended to fund many different types of projects around the state, two of the eight chapters in their subchapters contained language that could be inter- preted, in part, as serving an environmental justice or urban-focused agenda by either specifi- cally prioritizing “disadvantaged” (low-income) communities, communities with pollution burdens, or those undergoing population growth. The chapters did so, however, with language that differed in its specificity. In particular, Chapter 2 for “Safe Drinking Water and Water Quality Projects” directed $1.18 billion towards water quality projects with priority given to “projects that address chemi- cal and nitrate contaminants, other health hazards and by whether the community is disadvan- taged or severely disadvantaged.” [“Disadvantaged communities” have median household incomes less than 80% of the statewide average, “severely disadvantaged communities” less than 60%.] [16]. Chapter 2 prioritized communities along six criteria, including those stated above, and stated that at least one must be met. Chapter 8 for “Parks and Nature Education Facilities” stated: “The Department of Parks and Recreation shall include the following goals in setting spending priorities . . . The expan- sion of the state park system to reflect the growing population and shifting population centers and needs of the state” [16]. Chapter 9 for “Sustainable Communities and Climate Change Reduction” projects stated: “Acquisition and development of new parks and expansion of overused parks and recreation areas that provide park and recreational access to underserved communities shall be given preference.” And “creation of parks in neighborhoods where none currently exist shall be given preference” [16]. This section was enabled with more specific criteria through AB 31, the “Statewide Park Development and Community Revitalization Act of 2008”, which further directed funding for “the acquisition and development of parks and recreation areas and facili- ties in the communities that are currently least served by park and recreation facilities by PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 3 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure emphasizing the creation of park space and recreational opportunities and the expansion of park accessibility to underserved communities.” [A “critically underserved community” has <3 acres of usable parkland per 1,000 residents or is “disadvantaged” (see above)] [17]. For AB 31, the process of awarding these competitive grants was overseen by the Depart- ment of Parks and Recreation with quantitative criteria written by politicians and “diverse allies,” including Los Angeles social justice organization The City Project [18]. In their applica- tion guide, the department included a scoring rubric that awarded more points for new parks in areas where there are no existing parks, for applicants holding meetings to gather feedback from nearby residents, for being situated in critically underserved communities, and other detailed criteria [19]. There was also no requirement to match funds which might otherwise have put communities with less fiscal capacity at a disadvantage. The location of all local grants for the three prioritized chapters are mapped in Fig 1A–1D. Methods To assess how Proposition 84 funds were spent with regard to need and equity, we analyzed Proposition 84 spending at the level of census block groups (ranging in area from 0.015 to 16,000 km ). The underlying database was compiled by GreenInfo Network using grant infor- mation from the California Natural Resources Agency. This database included 2,152 projects for which a local footprint could be identified, thus enabling an analysis of the communities that benefited. Planning, technology, and large regional grants where a distinct local impact could not be identified were not included in this analysis, because we could not assign them to particular census blocks. Such projects include the construction of a website portal for the Cal- ifornia Stormwater Quality Association, the creation of an urban greening plan for the entire city of San Diego, and a construction feasibility study for Madera County. Large regional water projects, and all projects funded under Chapter 4, “Statewide Water Planning and Design” were also considered non-local and thus excluded from our analysis. While some of these grants provide local benefits, it was impractical to verify where these benefits accrued from the information provided, in contrast to grants where a particular location for a project was identi- fied. Grant funds were administered by 17 different agencies using different procedures and guidelines for soliciting and selecting projects for funding. While we were able to access a com- prehensive database of projects that were funded, there is no comprehensive records of proj- ects that were not funded. Therefore, we focused our analysis only on funded projects. Altogether, the 2,152 projects with local impacts that we analyzed accounted for $2 billion of the total spending under Proposition 84. We added data concerning several pertinent envi- ronmental and socioeconomic attributes associated with each census block group (see Table 2). These quantitative attributes allow us to examine whether California block groups that benefited from Proposition 84 projects differed from those that did not. Altogether, there are 23,212 block groups in California, of which 1,242 received Proposition 84 project funding with an identifiable local impact. We used environmental data from CalEnviroScreen (CES), a California-wide tool to assess at-risk communities. However, we did not use the entire composite CES index score, because it contains a large array of sub-indicators not relevant for this study. Instead, we used only the pollution burden components, which include measures for drinking water and potential groundwater contamination, impaired water bodies, levels of ozone and particulate matter, pesticide use, hazardous waste and cleanup sites, toxic releases, and traffic density. The result is a ranking of block groups according to pollution burdens that could potentially harm the health of residents. PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 4 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Fig 1. Location of all grants determined to have a local impact. Size corresponds to grant amount. Grants are heavily concentrated in population centers like Los Angeles and the San Francisco Bay Area, but were also disbursed throughout the Central Valley, Sierra Nevada, and along the coast. Geographic data from U.S. Census Bureau [20]. https://doi.org/10.1371/journal.pone.0211925.g001 Using the data in Table 2, we asked whether funding favored areas based on socioeconomic characteristics, urban demographics, a shortage of park space per capita, or environmental pol- lution burdens. We did this analysis for each chapter separately since the chapters differed in their intent, language, and the specificity of guidance regarding priorities. The analysis itself was conducted in two ways. First, we used hurdle models to see if the variables in Table 2 could predict receipt of grants or amount of funding. The Proposition 84 spending data con- tain a large number of zeroes because the vast majority (94%) of block groups did not receive PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 5 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Table 2. Variables and sources used in the analyses. Source Median household income American Community Survey 2009–13, U.S. Census Bureau [21] (MHI) Population density American Community Survey 2009–13, U.S. Census Bureau [22] Urban coverage TIGER 2013 Urban Areas, U.S. Census Bureau [23]. Coverage calculated as the percent of a block group that is covered by a Census urban area Park space per capita California Protected Areas Database 2015a, GreenInfo Network [24] CES pollution exposure CalEnviroScreen (CES) 2.0, California Office of Environmental Health Hazard measures Assessment [25] Population change 2000 Census of Population and Housing, U.S. Census Bureau; standardized to 2010 Census Bureau geography by the National Historical Geographic Information System [26] https://doi.org/10.1371/journal.pone.0211925.t002 any grants. While this presents problems for conventional linear models, we can model this type of data with a hurdle or two-part model. The “hurdle” is represented by a probit model that predicts the receipt of a grant based on median household income [27]. We use median household income as the predictor for grant receipt because successfully writing a grant pro- posal is linked to the resources available to a community which is likely reflected in household incomes [3,4]. The second part is a truncated linear regression model that predicts amount of funding conditional on passing the initial hurdle of receiving a grant. We model grant funding using predictor variables that capture need and environmental inequity. In addition to green space deficits and pollution, we include measures on population density and urban coverage to understand how well funding was disbursed to urban population centers. Taken together, the hurdle model allows us to evaluate how funding was awarded along attributes of need in Cali- fornia communities. The second part of the analysis was contrasting funded to unfunded block groups to evalu- ate how well all chapters funded low-income communities and how well the two parks chap- ters in particular (Chapters 8 and 9) funded park-poor urban communities. We used the same predictor variables for both Chapter 8 and Chapter 9, even though Chapter 8 did not specifi- cally prioritize low-income areas. This allowed us to compare how these two chapters served need, recognizing that the chapters differed in the specificity and implications of their lan- guage. For continuous variables where the distribution of sample means met the normality assumptions, we conducted a two-sample t-test contrasting funded to unfunded block groups; if there is no difference between the means of these two groups, then there is no evidence of targeting. For park-space per capita, which did not meet the assumptions necessary for a t-test, we conducted a chi-squared goodness of fit test between funded and unfunded block groups using binned frequency data. For each test, unfunded block groups were comprised of all block groups that did not receive funding from a given chapter, but may or may not have received funding from other chapters. Results In California, pollution burdens are negatively correlated with median household income such that higher income communities are slightly less likely to suffer high pollution burdens (Fig 2). Proposition 84 funding was distributed to block groups across all income levels. All chap- ters except Chapter 7 funded more grants in block groups at or below the average median household income than above, with Chapter 9 exhibiting the highest targeting of low-income block groups (Table 3). PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 6 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Fig 2. Median household income and CES pollution burden scores for all 23,312 California block groups (r = -0.29). Though low pollution burdens are present at all socioeconomic levels, greater pollution burdens are largely constrained to low-income block groups. https://doi.org/10.1371/journal.pone.0211925.g002 Across all block groups, those that received Proposition 84 funding had moderately lower incomes than unfunded block groups (Table 4). Chapters 2 and 9 stand out as being most Table 3. Funding and grants in block groups above and below average median household income. All chapters awarded more grants and funding to block groups below the average median household income, with the exception of Chapter 7. Chapter Spending and number of grants in block groups Spending and number of grants in block groups below average median household income above average median household income Chapter $105,123,711 $37,577,886 2 126 47 Chapter $154,049,689 $28,414,586 3 25 11 Chapter $349,716,208 $174,393,869 5 370 224 Chapter $224,204,698 $74,781,676 6 139 82 Chapter $67,036,158 $93,920,423 7 85 164 Chapter $127,090,114 $108,704,674 8 352 219 Chapter $378,631,864 $40,046,527 9 254 54 Total $1,405,852,442 $557,839,641 1,351 801 https://doi.org/10.1371/journal.pone.0211925.t003 PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 7 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Table 4. Mean median household income for funded and unfunded block groups, by chapter. Differences between funded and unfunded income was tested for significance using two-tailed t-tests. Chapters Average median household income in Average median household income in funded block groups unfunded block groups � � All Chapters $62,346 $68,203 � � Chapter 2 $59,811 $67,907 Chapter 3 NS NS Chapter 4 NS NS � � Chapter 5 $62,249 $67,838 Chapter 6 $62,772 $67,724 � � Chapter 7 $80,915 $67,529 Chapter 8 NS NS � � Chapter 9 $47,919 $68,080 p < 0.05 � � p < 0.01 NS identifies chapters for which the mean value for funded block groups did not differ significantly from the mean value of unfunded block groups https://doi.org/10.1371/journal.pone.0211925.t004 effective at funding those communities on the lower scale of income, whereas Chapter 7 funded higher income block groups on average. Chapter 7 funded coastal protection projects, but within the subset of coastal block groups (n = 676), funded block groups had lower median household income on average ($71,856) compared to unfunded block groups ($84,748), though still higher than the average for all block groups within the state ($67,690). For Chap- ters 3, 4 and 8, there was no significant income difference between funded and unfunded block groups. With clear differences in the pollution burdens of high and low-income block groups, and in the incomes of funded and unfunded block groups, we used a hurdle model to see if income could predict receipt of grants, and if population density, urban coverage, park space per cap- ita, and pollution burdens could predict funding amount conditional on receipt. The hurdle model revealed biases that were significant but had very small effect sizes. (Table 5). The prob- ability of a block group receiving a grant from any chapter decreased slightly with median household income, with block groups of $50,000 median income having a 1.5% greater proba- bility of receiving a grant than those of $150,000. The linear models yielded some small effects of the covariates on grant size. In particular, greater population density predicted slightly more funding from all grants, and Chapter 2 grants in particular, and higher pollution scores were associated with more funding from Chapter 9 grants. For all other chapters, the hurdle model had poor predictive power with wide margins of error, and did not reveal significant effects on funding from the covariates (S1 Fig). Surprisingly, even park space per capita was not a strong predictor of funding amount for Chapter 9 for which park-poor communities were a priority. Finally, we examined the two park chapters of Proposition 84, Chapters 8 and 9, and com- pared how well they funded park-poor urban communities. Our analysis revealed several strong biases (Table 6). Grants for Chapter 8, “Parks and Nature Education Facilities,” were preferentially awarded to sparsely populated rural areas with dramatically more park space per capita than unfunded block groups. In contrast, grants for Chapter 9, “Sustainable Communi- ties and Climate Change Reduction” favored areas with very little park space per capita. Block groups that received these grants did not differ significantly from unfunded block groups in terms of urban coverage and population density, but because most block groups in California are urban and densely populated, this does not necessarily signify a departure from the PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 8 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Table 5. Hurdle model results. Coefficient are estimated for grant selection and for magnitude of funding. The top values are coefficient estimates, the bottom values are standard errors. Probability of receiving a grant (logit) All Chapters Chapter 2 Chapter 3 Chapter 5 Chapter 6 Chapter 7 Chapter 8 Chapter 9 MHHI ($10,000) -0.03 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1695.20 1255.46 1337.71 1,912.71 1,046.21 1616.42 1972.73 Magnitude of Funding (Gaussian) � � � Population Density 63.30 219.42 2,793.32 69.60 -564.64 54.66 199.08 30.78 57.3 57.31 2,074.70 306.51 2,229.26 97.67 140.86 17.12 Percent Urban -7,506.37 2,151.03 46,998.48 -14,878.96 -16,184.73 -3,941.43 3,166.20 478.63 3,125.85 3,125.85 71,003.42 11,268.48 18,901.60 3,159.94 3,709.65 4,147.30 � � Park Acres per 1000 People -0.39 - 0.83 -255.48 - 2.18 9.07 - 1.24 -0.01 -2.34 1.40 1.40 512.85 2.07 2.93 3.18 0.25 4.35 � � � Pollution Burden Score 18,3151.40 127,843.70 -165,852.10 299,866.70 -11,030.34 18,740.86 618.60 175,753.20 74,678.52 74,678.52 1,940,278.00 213,025.20 179,894.30 62,791.01 67,511.45 65,904.17 � � � p > 0.001 � � p > 0.01 p > 0.05 https://doi.org/10.1371/journal.pone.0211925.t005 chapter’s stated intent. Neither Chapter 8 nor Chapter 9 favored areas with greater population growth. Discussion This paper examined the disbursement of funds from a major California environmental ballot measure with respect to household income, pollution, and other variables related to urban environmental justice. We found that while a majority of the proposition’s chapters did on average fund projects in lower income areas, some did not. In particular, funding for coastal protection and state parks under Chapters 7 and 8 went to wealthier areas. Within the subset of only coastal block groups, Chapter 7 grants went to those with lower median household income on average, though still higher than the average across for block groups across the state. This is a function of real estate values–coastal block groups are generally wealthier and have higher property costs compared to inland block groups. On the other hand, Chapters 2 and 9, which had language clearly prioritizing low-income communities, did very well at tar- geting funding to low-income block groups. The hurdle models revealed that the probability of a block group receiving a grant decreased slightly with median household income. However, income could not predict receipt of grants from individual chapters, possibly because there were so few block groups funded by Table 6. Evaluating funding from Chapters 8 and 9 to urban communities. Values are expressed as the ratio between the average value of funded and unfunded block groups for a given variable. Chapter 9. Ratio of funded to unfunded block groups � � � Variables Population density Urban coverage Park space per capita† Population growth 2000–2010 Chapter 8: 0.15 0.35 25.19 NS “Parks and Nature Education Facilities” Chapter 9: NS NS 0.55 NS “Sustainable Communities and Climate Change Reduction” Significance tested with two-sample, two-tailed t-test † Significance tested with chi-squared test using binned frequency data NS identifies a variable where the mean value for block groups funded by a given chapter did not differ significantly from the mean value of unfunded block groups https://doi.org/10.1371/journal.pone.0211925.t006 PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 9 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure any given chapter compared to the number of unfunded block groups in the state. Given that a block group received a grant, very few of the covariates of interest could predict the size of that grant. In other words, larger grants did not seem to necessarily go to the areas that were the most urban, the most park-poor, or the most polluted. There were two exceptions to this– greater population density was associated with larger Chapter 2 grants and worse pollution was associated with larger Chapter 9 grants. Surprisingly, park space and income were not good predictors of funding for Chapter 9 grants which specifically targeted disadvantaged urban communities. Overall, the receipt and size of grants could not be predicted even with a two-part model, possibly because of the sheer number of unfunded block groups and other fac- tors that may be important to the grant award process. When we examined the two park chapters, Chapters 8 and 9, we found that Chapter 9 was exceptionally successful at meeting its stated priorities of funding park projects in lower- income urban communities that lack park space. Chapter 8, though not intended to fulfill the same role as Chapter 9 projects, nevertheless had a stated priority to reflect growing population centers. This goal of targeting areas of population growth was not met, and grants were awarded overwhelmingly to sparsely populated rural areas. It should be noted that acres per 1,000 residents and population density are blunt measures that may not tell the whole story. More detailed analyses that distinguished multi-family from single family housing, or that used actual satellite images, or that reflected the extent to which neighborhoods even wanted more parks would be the next steps in this research. A review of grant objectives reveals that nearly 77% of funds from Chapter 9 went towards construction of new parks, compared to less than 5% of Chapter 8 funds. Thus, the two parks chapters served different needs in different populations, despite both intending to expand parks in the state. While our hurdle model revealed that lower income block groups were more likely to be awarded grants overall, the mechanisms behind submitting and being approved for a grant were not investigated in this study. As previous studies have noted, com- petitive grants tend to favor communities with the resources to submit strong proposals, or any proposals at all [8,9]. These factors, in addition to the social and economic capital required to obtain space and approval for parks, are an area of exploration for future research. One question that we tried to answer in this paper was if the funding disbursement matched the intention of the ballot measure in the sections of the ballot that did state a preference for disadvantaged communities. There is also the broader question of how well the funds reme- died inequities in environmental amenities and environmental quality overall even if the lan- guage did not specify such a goal. An argument can and has been made that public investments in the environment, particularly urban sustainability or park projects, should pri- oritize low-income communities because those communities often bear greater pollution bur- dens and green space deficits and have less local fiscal capacity to fund projects [10,28,29]. Evaluating how well public environmental investments serve low-income communities is important not only for gauging the commitment to environmental justice, but also for inform- ing how future legislation and ballot measures might be written. In the case of Proposition 84, language might have been an important factor. Chapter 8 was written with the somewhat vague goal of expanding parks to reflect growing and shifting population centers, yet funding tended to go to rural areas with an abundance of park space, no significant population growth, and no significant difference in income from unfunded areas. In contrast, Chapter 2 priori- tized low-income communities amongst other criteria and awarded nearly 3 times as much grant money to block groups below the average median household income. And finally, Chap- ter 9, with the strictest language explicitly constraining funding to underserved park-poor or low-income) communities, awarded over 9 times as much funding to low-income block groups. PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 10 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure In addition to language, there are likely systemic reasons that disadvantaged communities were not funded more in some parts of Proposition 84. A critical caveat for any analysis attempting to evaluate awards from a competitive grant process is that there is rarely, if ever, a database of grant applications that were rejected. This leads one to ask if poorer communities received fewer grants because they simply did not submit grants, or perhaps their proposals were not as professionally prepared. It is essential that a data base of successful and unsuccess- ful proposals is maintained, because only with those data could one suggest more precise rem- edies to the problem of underserving communities with public funds. If communities are not submitting grants, efforts at greater education and outreach from public agencies about fund- ing opportunities or grant-writing workshops might be needed. For instance, the Department of Parks and Recreation held a number of technical workshops for writing AB 31 grants [19]. But if disbursement agencies are not awarding grants to these communities even if they apply, then a more transparent award process is needed with greater efforts made towards serving these areas. Proposition 84 attempted to serve many different needs when it came to environmental infrastructure, with mixed results when it came to serving the urban communities most in need. This is perhaps not surprising, since the individuals and organizations that wrote the proposition and funded the campaign for this ballot measure tended to represent traditional conservation organizations and did not include environmental justice groups, aside from AB 31. As previous research has shown, nonprofits often act as powerful political actors in deter- mining, through contributions, what benefits are received from ballot measures [30]. In partic- ular, the two top funders of Proposition 84 were The Nature Conservancy $3,549,920) and California Conservation Action Fund $1,574,074) [31]. The lesson for public ballot measures is clear: the “devil is in the details” of the text. Since equity historically has not been a focus of these conservation groups, it may come as little surprise that most chapters of Proposition 84 did not contain stronger and more precise language directing money toward lower income or communities deprived of nature opportunities. However, there is some indication that envi- ronmental initiative sponsors and legislators in California are increasingly following the model established by Chapter 9 and AB 31 in Proposition 84, and are explicitly directing funds to communities in most need [32,33]. Ultimately, we should expect money to flow to these areas of greatest need only if they are prioritized, given access to grant-writing opportunities, and including in the writing of the measure. Supporting information S1 File. Data used for the analysis. (CSV) S1 Fig. Predicted probabilities of receiving a grant. (PDF) S2 Fig. Predicted magnitude of grant funding (All Chapters and Chapters 2, 3, 5). Pre- dicted grant funding for selected variables holding all others constant. (PDF) S3 Fig. Predicted magnitude of grant funding (Chapters 6, 7, 8, 9). Predicted grant funding for selected variables holding all others constant. (PDF) PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 11 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Acknowledgments This research was funded by the Resources Legacy Fund and the Anthony and Jeanne Pritzker Family Foundation. Data on funding was compiled and mapped by the GreenInfo Network. Author Contributions Conceptualization: Ian P. Davies, Jon Christensen, Peter Kareiva. Data curation: Ian P. Davies. Formal analysis: Ian P. Davies. Funding acquisition: Jon Christensen, Peter Kareiva. Investigation: Ian P. Davies. Methodology: Ian P. Davies, Peter Kareiva. Project administration: Ian P. Davies. Resources: Peter Kareiva. Software: Ian P. Davies. Supervision: Jon Christensen, Peter Kareiva. Validation: Ian P. Davies. Visualization: Ian P. Davies. Writing – original draft: Ian P. Davies, Jon Christensen, Peter Kareiva. Writing – review & editing: Ian P. Davies, Jon Christensen, Peter Kareiva. References 1. Trust for Public Land. TPL LandVote Database [Internet]. 2015 [cited 4 May 2016]. Available: TPL Land- Vote Database 2. Scarlett L. Citizens Vote Green: Approve $29 Billion in Land and Water Funding at the Ballot Box | Con- servancy Talk. In: The Nature Conservancy [Internet]. 2014 [cited 7 Aug 2017]. Available: https://blog. nature.org/conservancy/2014/11/05/citizens-vote-green-approve-29-billion-in-land-and-water-funding- at-the-ballot-box/ 3. Wolch J, Wilson JP, Fehrenbach J. Parks and Park Funding in Los Angeles: An Equity-Mapping Analy- sis. Urban Geogr. Taylor & Francis Group; 2005; 26: 4–35. https://doi.org/10.2747/0272-3638.26.1.4 4. Walls M. Parks and Recreation in the United States: Local Park Systems [Internet]. Washington, DC; 2009. Available: http://www.mparks.org/Portals/0/Resource-Center/Justifying Parks and Recreation/ Economic Impact/ResourcesfortheFuture-PandRintheUS-LocalParks.pdf 5. Marvier M, Wong H. Resurrecting the conservation movement. J Environ Stud Sci. Springer-Verlag; 2012; 2: 291–295. https://doi.org/10.1007/s13412-012-0096-6 6. Sze J, Gambirazzio G, Karner A, Rowan D, London J, Niemeier D. Best in Show? Climate and Environ- mental Justice Policy in California. Environ Justice. Mary Ann Liebert, Inc. 140 Huguenot Street, 3rd Floor New Rochelle, NY 10801 USA; 2009; 2: 179–184. https://doi.org/10.1089/env.2009.0028 7. Villamagna AM, Mogollo ´ n B, Angermeier PL. Inequity in ecosystem service delivery: socioeconomic gaps in the public-private conservation network. Ecol Soc. The Resilience Alliance; 2017; 22: art36. https://doi.org/10.5751/ES-09021-220136 8. Sister C, Wolch J, Wilson J. Got green? addressing environmental justice in park provision. GeoJournal. Springer Netherlands; 2010; 75: 229–248. https://doi.org/10.1007/s10708-009-9303-8 9. Loukaitou-Sideris A. Urban Form and Social Context: Cultural Differentiation in the Uses of Urban Parks. J Plan Educ Res. Sage PublicationsSage CA: Thousand Oaks, CA; 1995; 14: 89–102. https:// doi.org/10.1177/0739456X9501400202 PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 12 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure 10. Joassart-Marcelli P. Leveling the Playing Field? Urban Disparities in Funding for Local Parks and Recreation in the Los Angeles Region. Environ Plan A. SAGE PublicationsSage UK: London, England; 2010; 42: 1174–1192. https://doi.org/10.1068/a42198 11. Pincetl S. Nonprofits and Park Provision in Los Angeles: An Exploration of the Rise of Governance Approaches to the Provision of Local Services*. Soc Sci Q. Wiley/Blackwell (10.1111); 2003; 84: 979– 1001. https://doi.org/10.1046/j.0038-4941.2003.08404019.x 12. Garcı ´a R, Rawson Z, Yellott M, Zaldaña C. Healthy Parks, Schools and Communities for All: Park Development and Community Revitalization [Internet]. Los Angeles; 2009. Available: www. cityprojectca.org 13. Institute of Governmental Studies. Proposition 84 [Internet]. Berkeley, CA: University of California, Berkeley; 2006. Available: https://igs.berkeley.edu/library/elections/proposition-84 14. Smith-Heisters S, Summers AB. Analysis of California’s Propositions IE and 84: Funding the State’s Water and Flood Control Infrastructure [Internet]. Los Angeles; 2006. Available: http://reason.org/files/ 9c69395abc3615f6a55ef12355dcbd18.pdf 15. Lewis WS. Ballot-Box Environmentalism across the Golden State: How Geography Influences Califor- nia Voters’ Demand for Environmental Public Goods. Pomona Sr Theses. 2016; Available: https:// scholarship.claremont.edu/pomona_theses/149 16. Safe Drinking Water, Water Quality and Supply, Flood Control, River and Coastal Protection Bond Act of 2006 [Internet]. 84 2006 p. 14. Available: https://www.parks.ca.gov/pages/1008/files/prop_84_text. pdf 17. de Leo ´ n K. Statewide Park Development and Community Revitalization Act of 2008 [Internet]. 31 Sacra- mento, CA: California Assembly; 2008 p. 8. Available: https://www.parks.ca.gov/pages/1008/files/ab_ 31_bill_9-2008_chaptered.pdf 18. Park funds for park poor and income poor communities -Prop 84 and AB 31 standards are working! In: The City Project [Internet]. 2014 [cited 29 Oct 2018]. Available: https://www.cityprojectca.org/blog/ archives/32075 19. California Department of Parks and Recreation. Application Guide for the Statewide Park Development and Community Revitalization Program of 2008 [Internet]. Sacramento; 2009. Available: www.parks.ca. gov/grants. 20. United States Census Bureau. 2013 TIGER/Line Shapefiles [Internet]. 2013. Available: https://www. census.gov/cgi-bin/geo/shapefiles/index.php 21. United States Census Bureau. S1903 Median Income in the Past 12 Months (In 2013 Inflation-Adjusted Dollars) [Internet]. American FactFinder: 2009–2013 American Community Survey 5-Year Estimates: Demographic and Housing Estimates. 2013. Available: http://factfinder.census.gov 22. United States Census Bureau. B00001 Unweighted Sample Count of Total Population. In: American FactFinder: 2009–2013 American Community Survey 5-Year Estimates: Demographic and Housing Estimates [Internet]. 2013. Available: http://factfinder.census.gov 23. United States Census Bureau. 2013 Census Urban Areas. In: 2013 TIGER/Line Shapefile [Internet]. 2013. Available: https://www.census.gov/geo/maps-data/data/tiger.html 24. GreenInfo Network. California Protected Areas Data Portal 2015a [Internet]. 2015. Available: http:// www.calands.org/ 25. Office of Environmental Health Hazard Assessment. CalEnviroScreen Version 2.0 [Internet]. 2015 [cited 7 Aug 2017]. Available: https://oehha.ca.gov/calenviroscreen/report/calenviroscreen-version-20 26. Minnesota Population Center. National Historical Geographic Information System: Version 11.0 [Inter- net]. Minneapolis, MN: University of Minnesota; 2016. https://doi.org/10.18128/D050.V11.0 27. Henningsen A, Toomet O. maxLik: A package for maximum likelihood estimation in R. Comput Stat. Springer-Verlag; 2011; 26: 443–458. https://doi.org/10.1007/s00180-010-0217-1 28. Bullard RD. Environmental Justice in the 21st Century: Race Still Matters. Phylon (1960-). Clark Atlanta University; 2001; 49: 151. https://doi.org/10.2307/3132626 29. Heynen N, Perkins HA, Roy P. The Political Ecology of Uneven Urban Green Space The Impact of Polit- ical Economy on Race and Ethnicity in Producing Environmental Inequality in Milwaukee. Urban Aff Rev. 2006; 42: 3–25. https://doi.org/10.1177/1078087406290729 30. Pincetl S. Nonprofits and Park Provision in Los Angeles: An Exploration of the Rise of Governance Approaches to the Provision of Local Services*. Soc Sci Q. Blackwell Publishing; 2003; 84: 979–1001. https://doi.org/10.1046/j.0038-4941.2003.08404019.x 31. Cal-Access. Campaign finance—Water quality, safety and supply. Flood control. Natural resource pro- tection. Park improvements. Bonds. Initiative statute. In: California Secretary of State [Internet]. 2006 [cited 24 Aug 2017]. Available: http://cal-access.sos.ca.gov/Campaign/Measures/Detail.aspx?id= 1283864&session=2005 PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 13 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure 32. Rosenhall L. California Capitol focuses on environmental injustice—but will it lead to real results? In: CALmatters [Internet]. 2017 [cited 30 Aug 2017]. Available: https://calmatters.org/articles/california- capitol-hones-environmental-injustice-will-focus-lead-real-results/ 33. Megerian C. In the battle over California climate policies, green projects are now in the hot seat. Los Angeles Times. Los Angeles; 13 Mar 2017. Available: http://www.latimes.com/politics/la-pol-ca-offsets- environmental-justice-20170313-story.html PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 14 / 14 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png PLoS ONE Public Library of Science (PLoS) Journal

Assessing the flow to low-income urban areas of conservation and environmental funds approved by California’s Proposition 84

Loading next page...
1
 
/lp/public-library-of-science-plos-journal/assessing-the-flow-to-low-income-urban-areas-of-conservation-and-S15WlLY6zS
Publisher site
See Article on Publisher Site

Abstract

a1111111111 a1111111111 Government funding accounts for a large proportion of conservation and environmental improvements, and is often the result of citizen votes on state ballot measures. A key con- cern surrounding public investments in the environment is whether that funding serves lower-income communities, which are often the communities of greatest need. We applied OPENACCESS three statistical methods to analyze the spatial distribution of conservation funding derived from California’s Proposition 84, which distributed nearly $4 billion across California Citation: Davies IP, Christensen J, Kareiva P (2019) Assessing the flow to low-income urban between 2006 and 2015. First, we used hurdle models to ask if income, population density, areas of conservation and environmental funds urban coverage, or pollution could explain receipt of grants or magnitude of funding. Sec- approved by California’s Proposition 84. PLoS ONE ond, we compared the income levels of funded and unfunded communities for each chapter 14(2): e0211925. https://doi.org/10.1371/journal. of the proposition. Finally, we examine two sections of the proposition that were intended to pone.0211925 fund parks around the state and compare the attributes of funded and unfunded communi- Editor: Jacint Balaguer, Universitat Jaume I, SPAIN ties. Proposition 84 offers lessons for environmental legislation and future research. While Received: March 15, 2017 there were general tendencies for more funding to flow to poor areas and areas with pollu- Accepted: January 24, 2019 tion problems, the language in Proposition 84 as a whole was vague with respect to the Published: February 7, 2019 funding of disadvantaged areas, and as a result the targeting of these areas overall was at best modest. However, when enabling legislation (AB 31) defined specific “metrics of disad- Copyright:© 2019 Davies et al. This is an open access article distributed under the terms of the vantage” that had to be met by communities to receive funds from some sections of Proposi- Creative Commons Attribution License, which tion 84, the funds did flow much more selectively to poorer communities. This suggests that permits unrestricted use, distribution, and future ballot measures should be very explicit in their language if they want to promote equity reproduction in any medium, provided the original author and source are credited. in conservation investments, and that future research should investigate the extent to which technical workshops and outreach could further increase the number of funded grant pro- Data Availability Statement: All relevant data are within the paper and its Supporting Information posals from low-income communities. files. Funding: Jon Christensen received funding from Resources Legacy Fund, #11026 (http://www. resourceslegacyfund.org/). Ian P. Davies and Peter Kareiva received funding from the Anthony and Introduction Jeanne Pritzker Family Foundation. In the United States, ballot measures have become one of the largest funding mechanisms for Competing interests: The authors have declared that no competing interests exist. public investment in conservation and environmental improvements. In the last decade, $40 PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 1 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure billion of funding has been approved through ballot measures for conservation alone [1]. When a ballot measure or funding program is approved, headlines typically announce a vic- tory for the environment and a victory for conservation [2]. However, critics of these ballot measures have pointed out that the funding often fails to address social inequities in environ- mental protection and access to parks and natural areas [3,4]. Although minority groups such as Hispanics and African Americans consistently vote in favor of environmental measures, if these measures are seen to reinforce social inequality, this key support for conservation and environmental protection may be lost [5]. Thus it is increasingly important to examine conser- vation and environmental funding from an environmental justice perspective and ask how well low-income and high-need communities are served [6,7]. In their analysis of the 1996 Proposition K in Los Angeles, Wolch et al. found that park funding from this public ballot measure often compounded existing inequalities in park access by funding park improvements rather than investments in new properties [3]. Other studies suggest that funding often lacks a focus on structural inequity in park access; that is, urban residents tend to use park space more intensively than their rural and suburban counterparts, a fact that is ignored when measuring access by park space per unit area rather than per active user [3,8,9]. Following the money is important because if there are particular social or environmental needs that are not receiving funds, adjustments could be made for future policy. Yet to our knowledge, such a quantitative analysis has never been performed for a state ballot measure. Throughout California’s history, a number of ballot measures have been passed with the intention to fund environmental projects. These range from local measures like Proposition K in Los Angeles to the statewide Propositions 12 and 13 which all, in varying ways, intended to channel public funds to increasing park space and improving access. Minorities tend to live in cities with less local fiscal capacity to spend on parks, and for cities in California, public funds from state legislation and ballot measures have become a viable model for building infrastruc- ture [10]. Large environmental nonprofits play an important role in this process by helping to craft these measures through political partnerships and then donating to them so that they are passed [11]. While successful in passing park measures, some have expressed concern that nonprofits whose concern is habitat protection will prioritize green spaces on the edges of cit- ies, rather than in the urban core where few have access to open spaces [10]. When these deci- sions are made in writing measures and awarding grants, they can lead to a distribution of park resources that is not equitable for the communities most in need. An Overview of Proposition 84 Proposition 84, a general obligation bond, was passed in 2006 and at the time represented the largest state ballot measure in the United States for environmental protection. Notably, it was carried largely with support from California Latinos, who voted 84% in favor versus just 45% from non-Hispanic white voters [12]. Proposition 84 authorized $5.4 billion in spending on water quality and supply, natural resource protection, and urban greening in high-need areas– a wide breadth, leading some to criticize the measure for lacking clarity and accountability [13,14]. Others criticized the explicit earmarking of funds to specific groups, like the San Joa- quin River Conservancy, as evidence of too much sway from environmental donors [15]. Proposition 84 funded projects through a competitive grant process. Under “general provi- sions” the text of Proposition 84 specified the following social priorities: “assistance to commu- nities with contaminated sources of drinking water” and “revitalizing our communities and making them more sustainable by investing in . . . local parks and urban greening” [16]. The $5.4 billion was then divided by project type into nine distinct chapters, each with its own grant criteria for funding (Table 1). PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 2 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Table 1. Proposition 84 fund allocation from 2006–2015. Grants with local impacts are those projects we determined to have identifiable “on-the-ground” impacts in local communities, as opposed to planning and technology grants and large regional projects where local community impacts could not be identified. Chapter 1 of the bal- lot measure details general provisions of the proposition but has no funding tied to it, so it is excluded from this analysis. Chapters Funding Authorized Funding Awarded (in millions of dollars and number of grants) (millions) Grants with local Grants for which impacts are not readily assigned to All impact particular locations grants 2: Safe Drinking Water and Water Quality $1,525 $143 $1,003 $1146 Projects 173 796 969 3: Flood Control $800 $182 $411 $594 36 342 378 4: Statewide Water Planning and Design $65 $0 $64 $63 0 17 17 5: Protection of Rivers, Lakes, and Streams $928 $524 $132 $656 594 399 993 6: Forest and Wildlife Conservation $450 $299 $26 $325 221 30 251 7: Protection of Beaches, Bays, and Coastal $540 $161 $140 $301 Waters 249 266 515 8: Parks and Nature Education Facilities $500 $236 $67 $303 571 118 689 9: Sustainable Communities and Climate $580 $419 $102 $521 Change Reduction 308 256 564 Total $5,388 $2,152 $1,946 $3,909 2152 2224 4376 https://doi.org/10.1371/journal.pone.0211925.t001 While Proposition 84 was intended to fund many different types of projects around the state, two of the eight chapters in their subchapters contained language that could be inter- preted, in part, as serving an environmental justice or urban-focused agenda by either specifi- cally prioritizing “disadvantaged” (low-income) communities, communities with pollution burdens, or those undergoing population growth. The chapters did so, however, with language that differed in its specificity. In particular, Chapter 2 for “Safe Drinking Water and Water Quality Projects” directed $1.18 billion towards water quality projects with priority given to “projects that address chemi- cal and nitrate contaminants, other health hazards and by whether the community is disadvan- taged or severely disadvantaged.” [“Disadvantaged communities” have median household incomes less than 80% of the statewide average, “severely disadvantaged communities” less than 60%.] [16]. Chapter 2 prioritized communities along six criteria, including those stated above, and stated that at least one must be met. Chapter 8 for “Parks and Nature Education Facilities” stated: “The Department of Parks and Recreation shall include the following goals in setting spending priorities . . . The expan- sion of the state park system to reflect the growing population and shifting population centers and needs of the state” [16]. Chapter 9 for “Sustainable Communities and Climate Change Reduction” projects stated: “Acquisition and development of new parks and expansion of overused parks and recreation areas that provide park and recreational access to underserved communities shall be given preference.” And “creation of parks in neighborhoods where none currently exist shall be given preference” [16]. This section was enabled with more specific criteria through AB 31, the “Statewide Park Development and Community Revitalization Act of 2008”, which further directed funding for “the acquisition and development of parks and recreation areas and facili- ties in the communities that are currently least served by park and recreation facilities by PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 3 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure emphasizing the creation of park space and recreational opportunities and the expansion of park accessibility to underserved communities.” [A “critically underserved community” has <3 acres of usable parkland per 1,000 residents or is “disadvantaged” (see above)] [17]. For AB 31, the process of awarding these competitive grants was overseen by the Depart- ment of Parks and Recreation with quantitative criteria written by politicians and “diverse allies,” including Los Angeles social justice organization The City Project [18]. In their applica- tion guide, the department included a scoring rubric that awarded more points for new parks in areas where there are no existing parks, for applicants holding meetings to gather feedback from nearby residents, for being situated in critically underserved communities, and other detailed criteria [19]. There was also no requirement to match funds which might otherwise have put communities with less fiscal capacity at a disadvantage. The location of all local grants for the three prioritized chapters are mapped in Fig 1A–1D. Methods To assess how Proposition 84 funds were spent with regard to need and equity, we analyzed Proposition 84 spending at the level of census block groups (ranging in area from 0.015 to 16,000 km ). The underlying database was compiled by GreenInfo Network using grant infor- mation from the California Natural Resources Agency. This database included 2,152 projects for which a local footprint could be identified, thus enabling an analysis of the communities that benefited. Planning, technology, and large regional grants where a distinct local impact could not be identified were not included in this analysis, because we could not assign them to particular census blocks. Such projects include the construction of a website portal for the Cal- ifornia Stormwater Quality Association, the creation of an urban greening plan for the entire city of San Diego, and a construction feasibility study for Madera County. Large regional water projects, and all projects funded under Chapter 4, “Statewide Water Planning and Design” were also considered non-local and thus excluded from our analysis. While some of these grants provide local benefits, it was impractical to verify where these benefits accrued from the information provided, in contrast to grants where a particular location for a project was identi- fied. Grant funds were administered by 17 different agencies using different procedures and guidelines for soliciting and selecting projects for funding. While we were able to access a com- prehensive database of projects that were funded, there is no comprehensive records of proj- ects that were not funded. Therefore, we focused our analysis only on funded projects. Altogether, the 2,152 projects with local impacts that we analyzed accounted for $2 billion of the total spending under Proposition 84. We added data concerning several pertinent envi- ronmental and socioeconomic attributes associated with each census block group (see Table 2). These quantitative attributes allow us to examine whether California block groups that benefited from Proposition 84 projects differed from those that did not. Altogether, there are 23,212 block groups in California, of which 1,242 received Proposition 84 project funding with an identifiable local impact. We used environmental data from CalEnviroScreen (CES), a California-wide tool to assess at-risk communities. However, we did not use the entire composite CES index score, because it contains a large array of sub-indicators not relevant for this study. Instead, we used only the pollution burden components, which include measures for drinking water and potential groundwater contamination, impaired water bodies, levels of ozone and particulate matter, pesticide use, hazardous waste and cleanup sites, toxic releases, and traffic density. The result is a ranking of block groups according to pollution burdens that could potentially harm the health of residents. PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 4 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Fig 1. Location of all grants determined to have a local impact. Size corresponds to grant amount. Grants are heavily concentrated in population centers like Los Angeles and the San Francisco Bay Area, but were also disbursed throughout the Central Valley, Sierra Nevada, and along the coast. Geographic data from U.S. Census Bureau [20]. https://doi.org/10.1371/journal.pone.0211925.g001 Using the data in Table 2, we asked whether funding favored areas based on socioeconomic characteristics, urban demographics, a shortage of park space per capita, or environmental pol- lution burdens. We did this analysis for each chapter separately since the chapters differed in their intent, language, and the specificity of guidance regarding priorities. The analysis itself was conducted in two ways. First, we used hurdle models to see if the variables in Table 2 could predict receipt of grants or amount of funding. The Proposition 84 spending data con- tain a large number of zeroes because the vast majority (94%) of block groups did not receive PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 5 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Table 2. Variables and sources used in the analyses. Source Median household income American Community Survey 2009–13, U.S. Census Bureau [21] (MHI) Population density American Community Survey 2009–13, U.S. Census Bureau [22] Urban coverage TIGER 2013 Urban Areas, U.S. Census Bureau [23]. Coverage calculated as the percent of a block group that is covered by a Census urban area Park space per capita California Protected Areas Database 2015a, GreenInfo Network [24] CES pollution exposure CalEnviroScreen (CES) 2.0, California Office of Environmental Health Hazard measures Assessment [25] Population change 2000 Census of Population and Housing, U.S. Census Bureau; standardized to 2010 Census Bureau geography by the National Historical Geographic Information System [26] https://doi.org/10.1371/journal.pone.0211925.t002 any grants. While this presents problems for conventional linear models, we can model this type of data with a hurdle or two-part model. The “hurdle” is represented by a probit model that predicts the receipt of a grant based on median household income [27]. We use median household income as the predictor for grant receipt because successfully writing a grant pro- posal is linked to the resources available to a community which is likely reflected in household incomes [3,4]. The second part is a truncated linear regression model that predicts amount of funding conditional on passing the initial hurdle of receiving a grant. We model grant funding using predictor variables that capture need and environmental inequity. In addition to green space deficits and pollution, we include measures on population density and urban coverage to understand how well funding was disbursed to urban population centers. Taken together, the hurdle model allows us to evaluate how funding was awarded along attributes of need in Cali- fornia communities. The second part of the analysis was contrasting funded to unfunded block groups to evalu- ate how well all chapters funded low-income communities and how well the two parks chap- ters in particular (Chapters 8 and 9) funded park-poor urban communities. We used the same predictor variables for both Chapter 8 and Chapter 9, even though Chapter 8 did not specifi- cally prioritize low-income areas. This allowed us to compare how these two chapters served need, recognizing that the chapters differed in the specificity and implications of their lan- guage. For continuous variables where the distribution of sample means met the normality assumptions, we conducted a two-sample t-test contrasting funded to unfunded block groups; if there is no difference between the means of these two groups, then there is no evidence of targeting. For park-space per capita, which did not meet the assumptions necessary for a t-test, we conducted a chi-squared goodness of fit test between funded and unfunded block groups using binned frequency data. For each test, unfunded block groups were comprised of all block groups that did not receive funding from a given chapter, but may or may not have received funding from other chapters. Results In California, pollution burdens are negatively correlated with median household income such that higher income communities are slightly less likely to suffer high pollution burdens (Fig 2). Proposition 84 funding was distributed to block groups across all income levels. All chap- ters except Chapter 7 funded more grants in block groups at or below the average median household income than above, with Chapter 9 exhibiting the highest targeting of low-income block groups (Table 3). PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 6 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Fig 2. Median household income and CES pollution burden scores for all 23,312 California block groups (r = -0.29). Though low pollution burdens are present at all socioeconomic levels, greater pollution burdens are largely constrained to low-income block groups. https://doi.org/10.1371/journal.pone.0211925.g002 Across all block groups, those that received Proposition 84 funding had moderately lower incomes than unfunded block groups (Table 4). Chapters 2 and 9 stand out as being most Table 3. Funding and grants in block groups above and below average median household income. All chapters awarded more grants and funding to block groups below the average median household income, with the exception of Chapter 7. Chapter Spending and number of grants in block groups Spending and number of grants in block groups below average median household income above average median household income Chapter $105,123,711 $37,577,886 2 126 47 Chapter $154,049,689 $28,414,586 3 25 11 Chapter $349,716,208 $174,393,869 5 370 224 Chapter $224,204,698 $74,781,676 6 139 82 Chapter $67,036,158 $93,920,423 7 85 164 Chapter $127,090,114 $108,704,674 8 352 219 Chapter $378,631,864 $40,046,527 9 254 54 Total $1,405,852,442 $557,839,641 1,351 801 https://doi.org/10.1371/journal.pone.0211925.t003 PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 7 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Table 4. Mean median household income for funded and unfunded block groups, by chapter. Differences between funded and unfunded income was tested for significance using two-tailed t-tests. Chapters Average median household income in Average median household income in funded block groups unfunded block groups � � All Chapters $62,346 $68,203 � � Chapter 2 $59,811 $67,907 Chapter 3 NS NS Chapter 4 NS NS � � Chapter 5 $62,249 $67,838 Chapter 6 $62,772 $67,724 � � Chapter 7 $80,915 $67,529 Chapter 8 NS NS � � Chapter 9 $47,919 $68,080 p < 0.05 � � p < 0.01 NS identifies chapters for which the mean value for funded block groups did not differ significantly from the mean value of unfunded block groups https://doi.org/10.1371/journal.pone.0211925.t004 effective at funding those communities on the lower scale of income, whereas Chapter 7 funded higher income block groups on average. Chapter 7 funded coastal protection projects, but within the subset of coastal block groups (n = 676), funded block groups had lower median household income on average ($71,856) compared to unfunded block groups ($84,748), though still higher than the average for all block groups within the state ($67,690). For Chap- ters 3, 4 and 8, there was no significant income difference between funded and unfunded block groups. With clear differences in the pollution burdens of high and low-income block groups, and in the incomes of funded and unfunded block groups, we used a hurdle model to see if income could predict receipt of grants, and if population density, urban coverage, park space per cap- ita, and pollution burdens could predict funding amount conditional on receipt. The hurdle model revealed biases that were significant but had very small effect sizes. (Table 5). The prob- ability of a block group receiving a grant from any chapter decreased slightly with median household income, with block groups of $50,000 median income having a 1.5% greater proba- bility of receiving a grant than those of $150,000. The linear models yielded some small effects of the covariates on grant size. In particular, greater population density predicted slightly more funding from all grants, and Chapter 2 grants in particular, and higher pollution scores were associated with more funding from Chapter 9 grants. For all other chapters, the hurdle model had poor predictive power with wide margins of error, and did not reveal significant effects on funding from the covariates (S1 Fig). Surprisingly, even park space per capita was not a strong predictor of funding amount for Chapter 9 for which park-poor communities were a priority. Finally, we examined the two park chapters of Proposition 84, Chapters 8 and 9, and com- pared how well they funded park-poor urban communities. Our analysis revealed several strong biases (Table 6). Grants for Chapter 8, “Parks and Nature Education Facilities,” were preferentially awarded to sparsely populated rural areas with dramatically more park space per capita than unfunded block groups. In contrast, grants for Chapter 9, “Sustainable Communi- ties and Climate Change Reduction” favored areas with very little park space per capita. Block groups that received these grants did not differ significantly from unfunded block groups in terms of urban coverage and population density, but because most block groups in California are urban and densely populated, this does not necessarily signify a departure from the PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 8 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Table 5. Hurdle model results. Coefficient are estimated for grant selection and for magnitude of funding. The top values are coefficient estimates, the bottom values are standard errors. Probability of receiving a grant (logit) All Chapters Chapter 2 Chapter 3 Chapter 5 Chapter 6 Chapter 7 Chapter 8 Chapter 9 MHHI ($10,000) -0.03 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1695.20 1255.46 1337.71 1,912.71 1,046.21 1616.42 1972.73 Magnitude of Funding (Gaussian) � � � Population Density 63.30 219.42 2,793.32 69.60 -564.64 54.66 199.08 30.78 57.3 57.31 2,074.70 306.51 2,229.26 97.67 140.86 17.12 Percent Urban -7,506.37 2,151.03 46,998.48 -14,878.96 -16,184.73 -3,941.43 3,166.20 478.63 3,125.85 3,125.85 71,003.42 11,268.48 18,901.60 3,159.94 3,709.65 4,147.30 � � Park Acres per 1000 People -0.39 - 0.83 -255.48 - 2.18 9.07 - 1.24 -0.01 -2.34 1.40 1.40 512.85 2.07 2.93 3.18 0.25 4.35 � � � Pollution Burden Score 18,3151.40 127,843.70 -165,852.10 299,866.70 -11,030.34 18,740.86 618.60 175,753.20 74,678.52 74,678.52 1,940,278.00 213,025.20 179,894.30 62,791.01 67,511.45 65,904.17 � � � p > 0.001 � � p > 0.01 p > 0.05 https://doi.org/10.1371/journal.pone.0211925.t005 chapter’s stated intent. Neither Chapter 8 nor Chapter 9 favored areas with greater population growth. Discussion This paper examined the disbursement of funds from a major California environmental ballot measure with respect to household income, pollution, and other variables related to urban environmental justice. We found that while a majority of the proposition’s chapters did on average fund projects in lower income areas, some did not. In particular, funding for coastal protection and state parks under Chapters 7 and 8 went to wealthier areas. Within the subset of only coastal block groups, Chapter 7 grants went to those with lower median household income on average, though still higher than the average across for block groups across the state. This is a function of real estate values–coastal block groups are generally wealthier and have higher property costs compared to inland block groups. On the other hand, Chapters 2 and 9, which had language clearly prioritizing low-income communities, did very well at tar- geting funding to low-income block groups. The hurdle models revealed that the probability of a block group receiving a grant decreased slightly with median household income. However, income could not predict receipt of grants from individual chapters, possibly because there were so few block groups funded by Table 6. Evaluating funding from Chapters 8 and 9 to urban communities. Values are expressed as the ratio between the average value of funded and unfunded block groups for a given variable. Chapter 9. Ratio of funded to unfunded block groups � � � Variables Population density Urban coverage Park space per capita† Population growth 2000–2010 Chapter 8: 0.15 0.35 25.19 NS “Parks and Nature Education Facilities” Chapter 9: NS NS 0.55 NS “Sustainable Communities and Climate Change Reduction” Significance tested with two-sample, two-tailed t-test † Significance tested with chi-squared test using binned frequency data NS identifies a variable where the mean value for block groups funded by a given chapter did not differ significantly from the mean value of unfunded block groups https://doi.org/10.1371/journal.pone.0211925.t006 PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 9 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure any given chapter compared to the number of unfunded block groups in the state. Given that a block group received a grant, very few of the covariates of interest could predict the size of that grant. In other words, larger grants did not seem to necessarily go to the areas that were the most urban, the most park-poor, or the most polluted. There were two exceptions to this– greater population density was associated with larger Chapter 2 grants and worse pollution was associated with larger Chapter 9 grants. Surprisingly, park space and income were not good predictors of funding for Chapter 9 grants which specifically targeted disadvantaged urban communities. Overall, the receipt and size of grants could not be predicted even with a two-part model, possibly because of the sheer number of unfunded block groups and other fac- tors that may be important to the grant award process. When we examined the two park chapters, Chapters 8 and 9, we found that Chapter 9 was exceptionally successful at meeting its stated priorities of funding park projects in lower- income urban communities that lack park space. Chapter 8, though not intended to fulfill the same role as Chapter 9 projects, nevertheless had a stated priority to reflect growing population centers. This goal of targeting areas of population growth was not met, and grants were awarded overwhelmingly to sparsely populated rural areas. It should be noted that acres per 1,000 residents and population density are blunt measures that may not tell the whole story. More detailed analyses that distinguished multi-family from single family housing, or that used actual satellite images, or that reflected the extent to which neighborhoods even wanted more parks would be the next steps in this research. A review of grant objectives reveals that nearly 77% of funds from Chapter 9 went towards construction of new parks, compared to less than 5% of Chapter 8 funds. Thus, the two parks chapters served different needs in different populations, despite both intending to expand parks in the state. While our hurdle model revealed that lower income block groups were more likely to be awarded grants overall, the mechanisms behind submitting and being approved for a grant were not investigated in this study. As previous studies have noted, com- petitive grants tend to favor communities with the resources to submit strong proposals, or any proposals at all [8,9]. These factors, in addition to the social and economic capital required to obtain space and approval for parks, are an area of exploration for future research. One question that we tried to answer in this paper was if the funding disbursement matched the intention of the ballot measure in the sections of the ballot that did state a preference for disadvantaged communities. There is also the broader question of how well the funds reme- died inequities in environmental amenities and environmental quality overall even if the lan- guage did not specify such a goal. An argument can and has been made that public investments in the environment, particularly urban sustainability or park projects, should pri- oritize low-income communities because those communities often bear greater pollution bur- dens and green space deficits and have less local fiscal capacity to fund projects [10,28,29]. Evaluating how well public environmental investments serve low-income communities is important not only for gauging the commitment to environmental justice, but also for inform- ing how future legislation and ballot measures might be written. In the case of Proposition 84, language might have been an important factor. Chapter 8 was written with the somewhat vague goal of expanding parks to reflect growing and shifting population centers, yet funding tended to go to rural areas with an abundance of park space, no significant population growth, and no significant difference in income from unfunded areas. In contrast, Chapter 2 priori- tized low-income communities amongst other criteria and awarded nearly 3 times as much grant money to block groups below the average median household income. And finally, Chap- ter 9, with the strictest language explicitly constraining funding to underserved park-poor or low-income) communities, awarded over 9 times as much funding to low-income block groups. PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 10 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure In addition to language, there are likely systemic reasons that disadvantaged communities were not funded more in some parts of Proposition 84. A critical caveat for any analysis attempting to evaluate awards from a competitive grant process is that there is rarely, if ever, a database of grant applications that were rejected. This leads one to ask if poorer communities received fewer grants because they simply did not submit grants, or perhaps their proposals were not as professionally prepared. It is essential that a data base of successful and unsuccess- ful proposals is maintained, because only with those data could one suggest more precise rem- edies to the problem of underserving communities with public funds. If communities are not submitting grants, efforts at greater education and outreach from public agencies about fund- ing opportunities or grant-writing workshops might be needed. For instance, the Department of Parks and Recreation held a number of technical workshops for writing AB 31 grants [19]. But if disbursement agencies are not awarding grants to these communities even if they apply, then a more transparent award process is needed with greater efforts made towards serving these areas. Proposition 84 attempted to serve many different needs when it came to environmental infrastructure, with mixed results when it came to serving the urban communities most in need. This is perhaps not surprising, since the individuals and organizations that wrote the proposition and funded the campaign for this ballot measure tended to represent traditional conservation organizations and did not include environmental justice groups, aside from AB 31. As previous research has shown, nonprofits often act as powerful political actors in deter- mining, through contributions, what benefits are received from ballot measures [30]. In partic- ular, the two top funders of Proposition 84 were The Nature Conservancy $3,549,920) and California Conservation Action Fund $1,574,074) [31]. The lesson for public ballot measures is clear: the “devil is in the details” of the text. Since equity historically has not been a focus of these conservation groups, it may come as little surprise that most chapters of Proposition 84 did not contain stronger and more precise language directing money toward lower income or communities deprived of nature opportunities. However, there is some indication that envi- ronmental initiative sponsors and legislators in California are increasingly following the model established by Chapter 9 and AB 31 in Proposition 84, and are explicitly directing funds to communities in most need [32,33]. Ultimately, we should expect money to flow to these areas of greatest need only if they are prioritized, given access to grant-writing opportunities, and including in the writing of the measure. Supporting information S1 File. Data used for the analysis. (CSV) S1 Fig. Predicted probabilities of receiving a grant. (PDF) S2 Fig. Predicted magnitude of grant funding (All Chapters and Chapters 2, 3, 5). Pre- dicted grant funding for selected variables holding all others constant. (PDF) S3 Fig. Predicted magnitude of grant funding (Chapters 6, 7, 8, 9). Predicted grant funding for selected variables holding all others constant. (PDF) PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 11 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure Acknowledgments This research was funded by the Resources Legacy Fund and the Anthony and Jeanne Pritzker Family Foundation. Data on funding was compiled and mapped by the GreenInfo Network. Author Contributions Conceptualization: Ian P. Davies, Jon Christensen, Peter Kareiva. Data curation: Ian P. Davies. Formal analysis: Ian P. Davies. Funding acquisition: Jon Christensen, Peter Kareiva. Investigation: Ian P. Davies. Methodology: Ian P. Davies, Peter Kareiva. Project administration: Ian P. Davies. Resources: Peter Kareiva. Software: Ian P. Davies. Supervision: Jon Christensen, Peter Kareiva. Validation: Ian P. Davies. Visualization: Ian P. Davies. Writing – original draft: Ian P. Davies, Jon Christensen, Peter Kareiva. Writing – review & editing: Ian P. Davies, Jon Christensen, Peter Kareiva. References 1. Trust for Public Land. TPL LandVote Database [Internet]. 2015 [cited 4 May 2016]. Available: TPL Land- Vote Database 2. Scarlett L. Citizens Vote Green: Approve $29 Billion in Land and Water Funding at the Ballot Box | Con- servancy Talk. In: The Nature Conservancy [Internet]. 2014 [cited 7 Aug 2017]. Available: https://blog. nature.org/conservancy/2014/11/05/citizens-vote-green-approve-29-billion-in-land-and-water-funding- at-the-ballot-box/ 3. Wolch J, Wilson JP, Fehrenbach J. Parks and Park Funding in Los Angeles: An Equity-Mapping Analy- sis. Urban Geogr. Taylor & Francis Group; 2005; 26: 4–35. https://doi.org/10.2747/0272-3638.26.1.4 4. Walls M. Parks and Recreation in the United States: Local Park Systems [Internet]. Washington, DC; 2009. Available: http://www.mparks.org/Portals/0/Resource-Center/Justifying Parks and Recreation/ Economic Impact/ResourcesfortheFuture-PandRintheUS-LocalParks.pdf 5. Marvier M, Wong H. Resurrecting the conservation movement. J Environ Stud Sci. Springer-Verlag; 2012; 2: 291–295. https://doi.org/10.1007/s13412-012-0096-6 6. Sze J, Gambirazzio G, Karner A, Rowan D, London J, Niemeier D. Best in Show? Climate and Environ- mental Justice Policy in California. Environ Justice. Mary Ann Liebert, Inc. 140 Huguenot Street, 3rd Floor New Rochelle, NY 10801 USA; 2009; 2: 179–184. https://doi.org/10.1089/env.2009.0028 7. Villamagna AM, Mogollo ´ n B, Angermeier PL. Inequity in ecosystem service delivery: socioeconomic gaps in the public-private conservation network. Ecol Soc. The Resilience Alliance; 2017; 22: art36. https://doi.org/10.5751/ES-09021-220136 8. Sister C, Wolch J, Wilson J. Got green? addressing environmental justice in park provision. GeoJournal. Springer Netherlands; 2010; 75: 229–248. https://doi.org/10.1007/s10708-009-9303-8 9. Loukaitou-Sideris A. Urban Form and Social Context: Cultural Differentiation in the Uses of Urban Parks. J Plan Educ Res. Sage PublicationsSage CA: Thousand Oaks, CA; 1995; 14: 89–102. https:// doi.org/10.1177/0739456X9501400202 PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 12 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure 10. Joassart-Marcelli P. Leveling the Playing Field? Urban Disparities in Funding for Local Parks and Recreation in the Los Angeles Region. Environ Plan A. SAGE PublicationsSage UK: London, England; 2010; 42: 1174–1192. https://doi.org/10.1068/a42198 11. Pincetl S. Nonprofits and Park Provision in Los Angeles: An Exploration of the Rise of Governance Approaches to the Provision of Local Services*. Soc Sci Q. Wiley/Blackwell (10.1111); 2003; 84: 979– 1001. https://doi.org/10.1046/j.0038-4941.2003.08404019.x 12. Garcı ´a R, Rawson Z, Yellott M, Zaldaña C. Healthy Parks, Schools and Communities for All: Park Development and Community Revitalization [Internet]. Los Angeles; 2009. Available: www. cityprojectca.org 13. Institute of Governmental Studies. Proposition 84 [Internet]. Berkeley, CA: University of California, Berkeley; 2006. Available: https://igs.berkeley.edu/library/elections/proposition-84 14. Smith-Heisters S, Summers AB. Analysis of California’s Propositions IE and 84: Funding the State’s Water and Flood Control Infrastructure [Internet]. Los Angeles; 2006. Available: http://reason.org/files/ 9c69395abc3615f6a55ef12355dcbd18.pdf 15. Lewis WS. Ballot-Box Environmentalism across the Golden State: How Geography Influences Califor- nia Voters’ Demand for Environmental Public Goods. Pomona Sr Theses. 2016; Available: https:// scholarship.claremont.edu/pomona_theses/149 16. Safe Drinking Water, Water Quality and Supply, Flood Control, River and Coastal Protection Bond Act of 2006 [Internet]. 84 2006 p. 14. Available: https://www.parks.ca.gov/pages/1008/files/prop_84_text. pdf 17. de Leo ´ n K. Statewide Park Development and Community Revitalization Act of 2008 [Internet]. 31 Sacra- mento, CA: California Assembly; 2008 p. 8. Available: https://www.parks.ca.gov/pages/1008/files/ab_ 31_bill_9-2008_chaptered.pdf 18. Park funds for park poor and income poor communities -Prop 84 and AB 31 standards are working! In: The City Project [Internet]. 2014 [cited 29 Oct 2018]. Available: https://www.cityprojectca.org/blog/ archives/32075 19. California Department of Parks and Recreation. Application Guide for the Statewide Park Development and Community Revitalization Program of 2008 [Internet]. Sacramento; 2009. Available: www.parks.ca. gov/grants. 20. United States Census Bureau. 2013 TIGER/Line Shapefiles [Internet]. 2013. Available: https://www. census.gov/cgi-bin/geo/shapefiles/index.php 21. United States Census Bureau. S1903 Median Income in the Past 12 Months (In 2013 Inflation-Adjusted Dollars) [Internet]. American FactFinder: 2009–2013 American Community Survey 5-Year Estimates: Demographic and Housing Estimates. 2013. Available: http://factfinder.census.gov 22. United States Census Bureau. B00001 Unweighted Sample Count of Total Population. In: American FactFinder: 2009–2013 American Community Survey 5-Year Estimates: Demographic and Housing Estimates [Internet]. 2013. Available: http://factfinder.census.gov 23. United States Census Bureau. 2013 Census Urban Areas. In: 2013 TIGER/Line Shapefile [Internet]. 2013. Available: https://www.census.gov/geo/maps-data/data/tiger.html 24. GreenInfo Network. California Protected Areas Data Portal 2015a [Internet]. 2015. Available: http:// www.calands.org/ 25. Office of Environmental Health Hazard Assessment. CalEnviroScreen Version 2.0 [Internet]. 2015 [cited 7 Aug 2017]. Available: https://oehha.ca.gov/calenviroscreen/report/calenviroscreen-version-20 26. Minnesota Population Center. National Historical Geographic Information System: Version 11.0 [Inter- net]. Minneapolis, MN: University of Minnesota; 2016. https://doi.org/10.18128/D050.V11.0 27. Henningsen A, Toomet O. maxLik: A package for maximum likelihood estimation in R. Comput Stat. Springer-Verlag; 2011; 26: 443–458. https://doi.org/10.1007/s00180-010-0217-1 28. Bullard RD. Environmental Justice in the 21st Century: Race Still Matters. Phylon (1960-). Clark Atlanta University; 2001; 49: 151. https://doi.org/10.2307/3132626 29. Heynen N, Perkins HA, Roy P. The Political Ecology of Uneven Urban Green Space The Impact of Polit- ical Economy on Race and Ethnicity in Producing Environmental Inequality in Milwaukee. Urban Aff Rev. 2006; 42: 3–25. https://doi.org/10.1177/1078087406290729 30. Pincetl S. Nonprofits and Park Provision in Los Angeles: An Exploration of the Rise of Governance Approaches to the Provision of Local Services*. Soc Sci Q. Blackwell Publishing; 2003; 84: 979–1001. https://doi.org/10.1046/j.0038-4941.2003.08404019.x 31. Cal-Access. Campaign finance—Water quality, safety and supply. Flood control. Natural resource pro- tection. Park improvements. Bonds. Initiative statute. In: California Secretary of State [Internet]. 2006 [cited 24 Aug 2017]. Available: http://cal-access.sos.ca.gov/Campaign/Measures/Detail.aspx?id= 1283864&session=2005 PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 13 / 14 Follow the money: Environmental justice and spending from a conservation ballot measure 32. Rosenhall L. California Capitol focuses on environmental injustice—but will it lead to real results? In: CALmatters [Internet]. 2017 [cited 30 Aug 2017]. Available: https://calmatters.org/articles/california- capitol-hones-environmental-injustice-will-focus-lead-real-results/ 33. Megerian C. In the battle over California climate policies, green projects are now in the hot seat. Los Angeles Times. Los Angeles; 13 Mar 2017. Available: http://www.latimes.com/politics/la-pol-ca-offsets- environmental-justice-20170313-story.html PLOS ONE | https://doi.org/10.1371/journal.pone.0211925 February 7, 2019 14 / 14

Journal

PLoS ONEPublic Library of Science (PLoS) Journal

Published: Feb 7, 2019

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create folders to
organize your research

Export folders, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off