Research

INTRODUCTION

Nature and all who depend upon it are facing overlapping biodiversity and climate crises (Pachauri et al. 2014, Brondizio et al. 2019). Interest in environmental protection has increased, in response, as have actions by state and non-state actors. Protected areas (PAs) are a leading conservation response, mostly by governments. Well-designed and managed PAs can safeguard biodiversity (Gray et al. 2016) and contribute to climate mitigation (Duncanson et al. 2023). Protected areas do not always reduce economic pressures (Ferraro et al. 2013, Geldmann et al. 2013, Robalino et al. 2017, Shah et al. 2021), but they remain a conservation cornerstone that is complemented by other measures as many actors try to implement a proposed “30 x 30” target of the Convention on Biological Diversity’s post-2020 global framework, while also advancing protection efforts nationally (CBD 2022, U.S. DOI 2021).

However, PAs might not be permanent. Protected area downgrading, downsizing, and degazettement (PADDD) events are legal changes to reduce the type, number, and extent of activities in PAs, PAs’ boundaries, or the number of PAs, respectively (Mascia and Pailler 2011). At least 3803 such PADDD events have been enacted in 73 countries since 1892, affecting 5,971,562 km², alongside 1159 proposals in 26 countries (Golden Kroner et al. 2019, CI and WWF 2021). Certain PADDD events restored land or resource rights to local communities, alleviating prior conflicts due to PA creation (Andrade and Rhodes 2012) and supporting communities, while having no ecological impact. For example, in 2016, a PADDD enabled local Indigenous tribes to harvest plants in PAs (NPS 2016). This can avoid adverse impact on biodiversity and restore relationships between communities that were severed during colonization. However, most PADDD events (62%) are linked to industrial-scale resource extraction and economic development related to infrastructure industrial agriculture, mining, or oil/gas development (Golden Kroner et al. 2019, Naughton-Treves and Holland 2019). This can have negative impacts on habitat and biodiversity (Golden Kroner et al. 2016). A majority (82%) of PADDD events are recent (2000–2020), suggesting mounting conservation concerns.

Protected area downgrading, downsizing, and degazettement arise from conflicts about PAs. Myriad factors drive their spatial and temporal incidence, across countries or within one landscape. Several studies link PAs’ characteristics, i.e., distances, infrastructure, or slope, with PADDD risk (Tesfaw et al. 2018, Keles et al. 2020). However, results concerning such fixed factors leave unanswered questions concerning dynamics, including shifts in political representation from elections. The importance of elections in environmental policy is emphasized by List and Sturm (2006) and others. To our knowledge, political representation has never been analyzed in the context of PADDD and not within the Global North.

Study of the political economy of deforestation and environmental protection has demonstrated important roles for decentralization (Burgess et al. 2012) and elections (Pfaff et al. 2017, Pailler 2018, Ruggiero et al. 2021), in light of incentives to convert tropical forests to agricultural lands (Cisneros et al. 2021). Regarding governance, Bareille et al. (2023) emphasized potential impacts of democratization on PA implementation. The importance of ideology has also been highlighted (Chupp 2011). At a global level, Kammerlander and Schulze (2021) found centrist governments achieved better environmental performance than left or right oriented governments. For the U.S., research highlights roles of environmental and economic shocks within environmental voting (Herrnstadt and Muehlegger 2014, Elliot et al. 2023) and strategic behaviors by policy makers facing re-election (Brunell and Cease 2019), underlining the influence of competition, lobbies, and partisanship (McAlexander and Urpelainen 2020, Schulze 2021). We contribute to related literature by analyzing whether shifts in political representation influence the risks of PADDD.

Specifically, we examine how shifts in political representation affect the frequency of PADDD in political districts in the U.S., controlling for fixed differences through the use of a panel data set. The U.S. has significant biodiversity (Mittermeier and Mittermeier 1997) and is an historic leader in PAs (Richards 2018), though notably a recent PADDD hotspot (Golden Kroner et al. 2019). We use an exhaustive sample of PADDD events in terrestrial federally protected areas in the U.S., 2001 to 2018, to study the political economy of PADDD. The League of Conservation Voters (LCV) has a scorecard for elected officials’ conservation-policy votes that shows how parties can be polarized regarding conservation issues. For 2000–2020, Republican (Democratic) Senators/Representatives averaged 14% (82%)/12% (85%). We show that shifting party in power also affects risks of PADDD. Shifting House Representative to Republican increases PADDD risk in any district, especially for proposals. Protected area downgrading, downsizing, and degazettement is also more likely if there is a shift from Democratic to Republican in the majorities for the House or Senate.

METHODS

Data

Units of observation

Our aim is to analyze the impacts of changes in political variables on the probability of PADDD. Most PADDD events are proposed as bills by members of Congress before being enacted, i.e., passed into law. From PADDDtracker, e.g., “Representative Tom McClintock introduced legislation (H. R. 934) that would roll back the boundary of the Wild and Scenic Merced River in order to allow Merced Irrigation District to increase the height of New Exchequer Dam.” Protected area downgrading, downsizing, and degazettement events can also be directly enacted at the executive level by federal agencies through a process called rulemaking. Congressional elections occur every two years. The House of Representatives is made of 435 members, each representing a political district in a state, and each having to be re-elected every 2 years. The Senate is composed of 100 members, 2 for each of the 50 states, who serve 6-year terms such that ~1/3 must run again in each election.

As explained in Appendix 1, we focus on protected areas managed by federal agencies that meet the international definition of a PA (Dudley 2008) as in Golden Kroner et al. (2019). For those protected areas, PADDD events are either congressional decisions (i.e., legislation) or executive decisions (e.g., executive orders or regulations through federal agencies). Legislation is proposed by a member of Congress. Any proposal undergoes review and potential modifications by a committee before being voted on by the chamber. If approved by both chambers and signed by the president, the PADDD is enacted. Executive orders are ultimately decided upon by the President. Regulations are promulgated by federal agencies, whose leaders are appointed by the President. Specific processes to promulgate such regulations vary by agency but usually involve a proposal, public comment, and finalization.

To estimate the impact of elections on reductions in protection, we restrict our study to land within federal PAs in 2001 (PAD-US, USGS 2018). We overlapped each PA’s boundaries with districts’ boundaries. Because some districts shifted during 2001–2018, we construct spatial units of observation by intersecting all successive shapefiles for district boundaries (Lewis et al. 2013). Thus, a PA in only one district is considered as one unit while a PA in two districts, 1 and 2, is two units; then, if during our period a share of district 1 is redefined as being in district 2, a 3^rd unit is distinguished. This helps ensure units do not split during our period of analysis. We know in which district a protected area is in each year. Finally, we drop all resulting observations under 100 ha, due mainly to mis-overlaps.

Our panel for the period 2001–2018 has 1413 observations. Each unit is a portion of one of 435 U.S. districts that was in a protected area in 2001. On average, each district has just over three such spatially distinct sub-districts which, for us, function as distinct units of observation. For each of those observations, we can determine whether a PADDD event occurred in the area at any point in time, as well as identify the elected congressmen then, at the district and state levels.

PADDD: dependent variable

We use the available data on U.S. PADDD events from PADDDtracker (CI and WWF 2021), updated with refinements from Olsson et al. (2021). We focus on the 97% of post-2000 PADDD events for which polygons are available to be able to spatially locate them in our observational units. Thus, we can locate PADDD events in districts. We consider any attempt to modify PAs’ sizes or characteristics, both modifications already enacted and those just proposed. Our database includes 1094 PADDD events between 2001 and 2018 (233 enactments and 861 proposals), as seen over time in Figure 1 and over space in Figure 2 (with more details on the PA and PADDD data in Appendix 1).

Figure 1 reveals clusters of PADDD events in specific years, highlighting that a single decision can result in multiple PADDD events, i.e., for more than one PA, perhaps even throughout the U.S. Approximately 95% of events exhibit this characteristic.

Protected area downgrading, downsizing, and degazettement events list proximate causes, i.e., reasons why events are proposed or enacted. Table 1 summarizes the listed proximate causes for the post-2000 PADDD events that we study. Most of the non-enacted proposals are related to infrastructure, others to mining, or to oil-and-gas, whereas the enacted events mostly list subsistence. For example, restoring rights for local tribes to harvest plants; notably these events are not a conservation concern for biodiversity conservation because they specify that the authorization can only be made if there is no ecological impact, and also because they restore rights to Indigenous peoples, offering positive cultural and social values. Other listed causes for these U.S. PADDD events include access for snowmobiles, off-road vehicles, hovercrafts for hunting, grazing, backcountry stock, and paddling. Some events are proposed but not enacted.

In our data, each proposal or enactment is an event. If the proposal is enacted, it becomes a new event. Our dependent variable is that PADDD has been proposed or enacted in a given unit in a given year. Overall, 32% of our units experienced at least one PADDD event. Among those, 48% experienced PADDD only once. For 2% of units, PADDD occurred more than three times during 2001–2018. For most of the units with more than one event in total, the events did not happen in the same year.

Independent variables

We gathered complete information concerning the elected representatives’ political parties from the League of Conservation Voters (LCV 2021). In a given year, we know at the district level if the Representative is Republican and, at a state level, how many of the (two) Senators are Republicans.

We can also control for several potentially relevant economic conditions using data from the Bureau of Economic Analysis (US BEA 2020) at the county level. Then using the counties’ centroids, we place each county in a district to aggregate the data up to district level. We compute the district’s GDP, as well as the shares of GDP in land-intensive sectors: agriculture, mining, or oil and gas.

Empirical analysis

We estimate a fixed-effect model to identify the impacts of political dynamics, i.e., changes over time in political representation, on probabilities of PADDD. We developed the following equation:

(1)

In Equation 1, PADD_ijst dependent variable equals one if at least one PADDD event occurred in unit i in district j in state s in year t, else it equals zero. It equals 1 for 779 observations. Our dependent variable equals one even if multiple events occur during a year. This explains why our variable is equal to one for only 779 observations, despite 1094 total PADDD events in our sample, while its mean value is 0.031, and standard deviation is 0.172. D_jst equals one if the elected House Representative for district j in state s is Republican at time t. S_st equals one if at least one of the Senators in state s is Republican at time t. M_mt is a set of m = 2 variables for political majorities in the Senate and House of Representatives, each equal to one for years t in which Republicans have majorities. Control variables X_kjst include GDP and its shares in land-intensive sectors. W_nt are the Presidential time periods, equal to one for all the years in question (We assign G. W. Bush’s first term as the reference period to which all the other presidential time periods are compared).

In (1), ϑ_i are fixed effects to control for all the time-unvarying confounding factors at unit level. Those include the distance to nearest metropolis and average level of explanatory variables such as the share of the economy in land-intensive sectors (agriculture, mining, oil-and-gas sectors) or the average share of Republican voters. Our standard errors are always clustered at the state level. Given that some PADDD events result from a single decision, one might suggest clustering the results at the protected area level. However, we have chosen to cluster at the state level for two reasons. Firstly, decisions leading to multiple PADDDs can impact several protected areas based on their location (e.g., proximity to borders) rather than a single area. Second, Abadie et al. (2023) recommended clustering standard errors at the treatment level. Since our paper examines the impact of elections at the district and state level, it is appropriate to cluster the standard errors at the state level. Note that clustering at the state level is more conservative than clustering at a district level.

Our estimated coefficients are impacts of changes in political representation on risk of PADDD. Political shifts are common in our study period. For the House of Representatives, at district level the overall standard deviation is 0.49, with between and within variation of 0.39 and 0.31, respectively. For Senate, overall standard deviation is also 0.49 and the between and within variation are 0.39 and 0.30, respectively. M_mt and W_nt vary only over time. θ_n does not identify impacts of presidencies, per se, but helps to control for time trends, including political circumstances. Therefore, δ_m captures the impact of changes in majorities within any presidential period. There have been majority shifts during three out of the five presidential time periods, for both Senate and House, although occurring under different presidencies for those two bodies.

Equation (1) includes a comprehensive set of controls. With fixed effects for units of observation and presidential periods, coefficients capture the effects on PADDD of changes within a specific unit of observation and presidency. For instance, coefficient Β captures the impact of a change in the party affiliation of a specific district on the risk of experiencing a PADDD while holding the president constant. Similarly, coefficient δ_m captures the impact of a change in a majority in a political body on the risk of PADDD within specific units of areas, while holding the president fixed.

Despite extensive controls, a concern could arise if year-specific events impact both our drivers of interest (majorities and the parties of House and Senate incumbents) and the incidence of PADDD events. This is relevant because many PADDD events follow common decisions in specific years (Fig. 1). That said, our inclusion of each of the presidency periods as controls account for one timescale of specific events, i.e., every four years, which could affect our political variables and decisions to propose or enact PADDD in multiple areas.

Still, we added two kinds of robustness tests. First, we exclude 2011 from our regression analyses. Second, we include year fixed-effects in our estimation to control for potential year-specific events that may influence drivers of interest and occurrences of PADDD. The latter approach effectively accounts for time varying factors that may influence our results. However, it constrains our ability to estimate coefficients for drivers that vary solely over time. For instance, the inclusion of year effects prevents us from estimating the impact of majorities in political bodies. This is the rationale behind our decision to prioritize the specification presented in Equation (1) for our main findings.

We also add six alternative specifications and estimators in the supplementary materials (Appendix 2). First, we examine proposals alone. Second, we estimate the model using conditional logit, which excludes units that never had PADDD. Third, we use an estimator, developed by de Chaisemartin and D’Haultfœuille (2024), robust to switches in treatment status and heterogeneities in treatment effects. Fourth, we look at the size and number of PADDD. Fifth, we separate state and district level. Sixth, we focus on contested elections whose results are supposedly more random.

RESULTS

Table 2 presents our main estimations considering all PADDD events, with columns adding independent variables, noting that when R-squared rises and the value of information criterion falls. Column (1) presents our minimal specification, simply using district and state representation plus some controls. Columns (2) and (3), respectively, add majorities and controls for the presidencies.

Column (4) displays our favored specification, in which we include all variables in Equation 1. Explanatory power, as reflected by the R², remains relatively low, even within Column 4 (5%). However, these magnitudes are relatively common in micro-econometric analyses. Many factors can idiosyncratically affect PADDD, such as local or national political shocks or natural hazards.

Shifting to having a Republican as a district’s House Representative significantly and positively increases the risk of PADDD in that district, all else being equal. As we include fixed effects for units, we are finding that even though some decisions are formally made at the national level, if a district switches from Democratic to Republican, PADDD risk increases in that specific area. This result is consistent with the influence of lobbying by district representatives, within national institutions, or at least their importance if other politicians interpret having local Republicans as a favorable signal to propose PADDD in a district. However, this is not the case for local senators.

Robustness checks (Appendix 2) support our findings. First, including year fixed effects confirms our results for local determinants, while making it impossible to assess the effect of congressional majorities. Second, we also re-estimate this model using an estimator, developed by de Chaisemartin and D’Haultfœuille (2024), robust to switches in treatment status and heterogeneities in treatment effect. This also allows us to estimate the lagged impact of our variable of interest on the risk of PADDD and to demonstrate that the parallel trend assumption holds for our estimation. Our results are confirmed and suggest that proposals would particularly occur shortly after the election, in the next year, rather than in later years of the mandates. Third, we excluded observations from 2011 because there were multiple PADDD events that year, which could have introduced specific biases. This does not affect the overall pattern of our results.

Fourth, restricting our sample to proposals, i.e., essentially to non-subsistence-related events (as per Table 1), reveals that our results are driven by the proposals. It would be highly relevant to also closely examine the political factors that influenced these enacted PADDD events related to subsistence in the U.S. context. They primarily restored rights to Indigenous populations displaced by PA establishment. Their political dynamics are likely to differ significantly from those of non-subsistence PADDD events. However, approximately 80% of these events occur in the same year (2016), which poses a challenge to identification because we cannot control for potential confounding factors specific to that year. Consequently, we are unable to address this particular question.

Fifth, using conditional logit confirms our results while restricting the sample to units that at some point experienced PADDD. Sixth, we look at the impact of our explanatory variables on the number of PADDD events, as well as on their sizes, which both also confirm our results. Seventh, we run the estimations of Table 2 separately for the House Representative and the Senators to make sure that our results are not driven by multicollinearity. Finally, building on insights from Pacca et al. (2021) and others, we re-examine the impact of Senators and House Representative using only tight (close, highly contested) elections. Their outcomes are likely to be more random, allowing for the identification of an unbiased causal impact. Due to the limited numbers for both PADDD events and tight elections, in our sample, estimates cannot indicate significant impacts.

DISCUSSION

Unlike PADDD studies to date (e.g., Symes et al. 2016), which focused on the spatially varying opportunity costs of PAs as drivers, empirically using fixed proxies, we focus on the political shifts over time that are generated by elections. For the U.S., we find that shifts in the party in power matter. Specifically, Republican majorities and local Republican leaders increase risks of PADDD events.

Our results drive off PADDD proposals more than enactments. Enactments require considerably more legislative effort, which limits our ability to detect the influence of elections on those events. Although PADDD proposals do not always result in an enacted legal change, they are consequential and, more generally, they clearly represent pressures and signal political stances on environmental efforts. The PADDD proposals may affect political discourse and the perceived political acceptability of such changes, i.e., can move “the Overton window” of policies worth discussing (Russell 2006).

Many PADDD events may be correlated because multiple events can stem from one decision. Those triggering multiple PADDD events can impact extensive areas across multiple states. Such decisions carry considerable weight within our results, which reinforces the relevance and significance of our findings in the context of environmental conservation policies.

Two possible mechanisms that could underlie the results we have found complement each other. First, Republicans may be intrinsically less inclined toward environmental conservation, increasing their likelihood of proposing PADDD. Second, it is also possible that unobservable confounding factors, such as shocks to voter preferences toward the environment, increase PADDD and the election of Republican congressmen. This second story suggests that Republicans tend to propose PADDD events more easily as a response to any such shifts in voters’ preferences. Following either of those mechanisms, such election results contribute to a higher risk of PADDD.

It is noteworthy that we observed significant influence of local political factors, despite many of our events resulting from relatively few national decisions. This suggests that alignment of interests at local levels can create opportunities for proposals that then resonate at the national level. Conversely, it is likely PADDD is not proposed at the national level if local Congress members are against it. This interpretation aligns with our results for majorities in the House and Senate, highlighting the substantial influence of political circumstances on political discourse and resulting proposals.

Many PADDD events, both enacted and simply proposed, were observed during a Democratic administration, that of President Obama. At first glance, those seem contrary to our overall story about the political economy of PADDD. However, details for this set of events support our story. Most of these enacted events resulted from a regulation that enabled local tribes to harvest plants in national parks, under the condition that authorizations have no negative ecological impacts. They advanced social justice per restoration of rights to original inhabitants. In significant contrast, not enacted during this period were proposals advanced by Congress that attempted to open the Arctic National Wildlife Refuge to oil-and-gas development. Such downgrading did not pass until Trump’s administration. Finally, multiple proposals waived environmental laws on land 100 miles from the U.S. border, for national security-related infrastructure, but they too were not passed.

Other political mechanisms (executive orders by presidents, regulations, and court decisions) also play important roles in conservation policy. For instance, given that most of the American public supports environmental protection (https://news.gallup.com/poll/1615/environment.aspx, Gallup 2022), it is worth understanding why increases in new protection of land have stalled since 2000 (Richards 2018). The PADDD trends in the U.S. are also relevant to the Biden Administration’s commitment to expanding protected lands and waters to 30% by 2030 (US DOI 2021), maybe including recognition of governance beyond traditional PAs.

Our analyses found key roles for political dynamics in shaping conservation in the U.S., home to the world’s first modern PA (Yosemite Land Grant in 1864, Yellowstone National Park in 1872). National parks have been called America’s best idea and expanded by both Democratic and Republican Presidents during the 20^th and 21^st centuries. However, since 2000, there has been a notable shift, with 94% of PADDD proposals and enactments occurring during a period of heightened political polarization on environmental issues, as documented by Milman (2019) using data from the League of Conservation Voters. Such political tensions rose during the COVID-19 pandemic (Casola et al. 2022), alongside rollbacks of environmental policy within the U.S. and globally (Golden Kroner et al. 2021). Our results show that Democratic and Republican policy makers seemingly took up polarized views on environmental issues in the case of conservation policies.

Contextual elements including timing and causes of PADDD events are essential to understand. Early PAs in the U.S. involved displacement and eviction of Native peoples (Dowie 2009) and some recent PADDD subsistence events have restored tribes’ rights to natural resources, enhancing social justice without degrading conservation (Naughton-Treves and Holland 2019). However, most proposed events in our data (and globally) are for infrastructure, industrial-scale resource extraction, and development. Similar development-related PADDD events, largely in the tropics, have negatively impacted forests, challenging conservation efforts (Forrest et al. 2015, Golden Kroner et al. 2016, Pack et al. 2016, Keles et al. 2020, 2023).

Our results suggest non-subsistence rather than social-justice motivations dominated the two parties’ perceptions of conservation-development trade-offs, such that rationales other than such conservation orientation could drive negative environmental outcomes (Forrest et al. 2015). For example, many U.S. events pertained to security infrastructure near the border with Mexico. The construction of physical barriers across protected lands in these areas clearly would negatively affect conservation outcomes there, such as habitat loss and fragmentation (Peters et al. 2018).

Continued attention to the political economy of protection, along with alternative conservation approaches such as Indigenous, community-led, and private systems, is crucial (S. Qin, Y. He, R. E. Golden Kroner, S. Shrestha, B. H. Coutinho, M. Karmann, J. C. Ledezma, C. Martinez, V. Morón-Zambrano, R. Ulloa, E. Yerena, C. Bernard, J. W. Bull, E. Mendoza, N. de Pracontal, K. Reytar, P. Veit, C. L. Matallana‐Tobón, L. Alden Wily, and M. B. Mascia, unpublished manuscript). The ongoing occurrence of PADDD and other challenges in expanding conservation protection underscores the need for significant adjustments in economic, social, and political functioning, including in production and consumption, to mitigate humans’ impacts on nature (Brondizio et al. 2019).

Protected area downgrading, downsizing, and degazettement studies for other countries could also consider shifts in political representation and other political factors. Results in Pailler (2018) and Ruggiero et al. (2021) on effects of elections in Brazil, for instance, suggest extensions to tropical countries given the deforestation risks implied by PADDD (Keles et al. 2023). Moreover, per List and Sturm (2006) or Elliott et al. (2023), one could study how the intensity of competition in those elections, or natural shocks, affect impacts. Protected area downgrading, downsizing, and degazettement may also vary with governance including rule of law, political stability, and corruption.

CONCLUSION

We shed light on the effects on environmental policies of shifts in political representation. We reveal a strong association between PADDD events (in particular, proposals) and the political landscape. We focus on shifts to Republican representation in districts and Republican majorities in the House and Senate. Thus, the political affiliation of elected officials and the composition of legislative bodies help to shape proposals that could potentially undermine conservation efforts. We confirm that when considering the fate of PAs, alongside their opportunity costs (Tesfaw et al. 2018, Keles et al. 2020), changes in key political determinants can be very important factors.

LITERATURE CITED

Abadie, A., S. Athey, G. W. Imbens, and J. M. Wooldridge. 2023. When should you adjust standard errors for clustering? Quarterly Journal of Economics 138(1):1–35. https://doi.org/10.1093/qje/qjac038

Andrade, G. S. M., and J. R. Rhodes. 2012. Protected areas and local communities: an inevitable partnership toward successful conservation strategies? Ecology and Society 17(4):14. https://doi.org/10.5751/ES-05216-170414

Bareille, F., J. Wolfersberger, and M. Zavalloni. 2023. Institutions and conservation: the case of protected areas. Journal of Environmental Economics and Management 118:102768. https://doi.org/10.1016/j.jeem.2022.102768

Brondizio, E. S., J. Settele, S. Díaz, and H. T. Ngo. 2019. Global assessment report on biodiversity and ecosystem services of the Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services. IPBES Secretariat, Bonn, Germany. https://doi.org/10.5281/zenodo.3831673

Brunell, T. L., and B. Cease. 2019. How do state-level environmental policies impact the voting behavior of national legislators? Social Science Quarterly 100(1):289–306. https://doi.org/10.1111/ssqu.12559

Burgess, R., M. Hansen, B. A. Olken, P. Potapov, and S. Sieber. 2012. The political economy of deforestation in the tropics. Quarterly Journal of Economics 127(4):1707–1754. https://doi.org/10.3386/w17417

Casola, W. R., J. M. Beall, M. N. Peterson, L. R. Larson, S. B. Jackson, and K. T. Stevenson. 2022. Political polarization of conservation issues in the era of COVID-19: an examination of partisan perspectives and priorities in the United States. Journal for Nature Conservation 67:126176. https://doi.org/10.1016/j.jnc.2022.126176

Chupp, B. A. 2011. Environmental constituent interest, green electricity policies, and legislative voting. Journal of Environmental Economics and Management 62(2):254–266. https://doi.org/10.1016/j.jeem.2011.03.008

Cisneros, E., K. Kis-Katos, and N. Nuryartono. 2021. Palm oil and the politics of deforestation in Indonesia. Journal of Environmental Economics and Management 108:102453. https://doi.org/10.1016/j.jeem.2021.102453

Conservation International and World Wildlife Fund. 2021. PADDDtracker: tracking protected area downgrading, downsizing, and degazettement.

Convention on Biological Diversity. 2022. Kunming-Montreal global biodiversity framework. Convention on Biological Diversity, Montreal, Quebec, Canada. https://www.cbd.int/doc/decisions/cop-15/cop-15-dec-04-en.pdf

de Chaisemartin, C., and X. D’Haultfœuille. 2024. Difference-in-differences estimators of intertemporal treatment effects. Review of Economics and Statistics 1–45. https://doi.org/10.1162/rest_a_01414

Dowie, M. 2009. Conservation refugees: the hundred-year conflict between global conservation and native peoples. MIT Press, Cambridge, Massachusetts, USA. https://doi.org/10.7551/mitpress/7532.001.0001

Dudley, N. 2008. Guidelines for applying protected area management categories. IUCN, Gland, Switzerland. https://doi.org/10.2305/IUCN.CH.2008.PAPS.2.en

Duncanson, L., M. Liang, V. Leitold, J. Armston, S. M. Krishna Moorthy, R. Dubayah, S. Costedoat, B. J. Enquist, L. Fatoyinbo, S. J. Goetz, M. Gonzalez-Roglich, C. Merow, P. R. Roehrdanz, K. Tabor, and A. Zvoleff. 2023. The effectiveness of global protected areas for climate change mitigation. Nature Communications 14:2908. https://doi.org/10.1038/s41467-023-38073-9

Elliott, R. J. R., V. Nguyen-Tien, E. A. Strobl, and T. Tveit. 2023. Climate-related natural disasters and voting behavior: evidence from environmental legislation in the US Senate. Journal of the Association of Environmental and Resource Economists 10(3):753–786. https://doi.org/10.1086/722540

Ferraro, P. J., M. M. Hanauer, D. A. Miteva, G. J. Canavire-Bacarreza, S. K. Pattanayak, and K. R. E. Sims. 2013. More strictly protected areas are not necessarily more protective: evidence from Bolivia, Costa Rica, Indonesia, and Thailand. Environmental Research Letters 8(2):025011. https://doi.org/10.1088/1748-9326/8/2/025011

Forrest, J. L., M. B. Mascia, S. Pailler, S. Z. Abidin, M. D. Araujo, R. Krithivasan, and J. C. Riveros. 2015. Tropical deforestation and carbon emissions from protected area downgrading, downsizing, and degazettement (PADDD). Conservation Letters 8(3):153–161. https://doi.org/10.1111/conl.12144

Geldmann, J., M. Barnes, L. Coad, I. D. Craigie, M. Hockings, and N. D. Burgess. 2013. Effectiveness of terrestrial protected areas in reducing habitat loss and population declines. Biological Conservation 161:230–238. https://doi.org/10.1016/j.biocon.2013.02.018

Golden Kroner, R. E., E. B. Barbier, O. Chassot, S. Chaudhary, L. Cordova, Jr., A. Cruz-Trinidad, T. Cumming, J. Howard, C. Karibuhoye Said, Z. Kun, A. Ogena, F. Palla, R. S. Valiente, S. Troëng, A. Valverde, R. Wijethunga, and M. Wong. 2021. COVID-era policies and economic recovery plans: are governments building back better for protected and conserved areas. Parks 27(Special Issue):5. https://doi.org/10.2305/IUCN.CH.2021.PARKS-27-SIRGK.en

Golden Kroner, R. E., R. Krithivasan, and M. B. Mascia. 2016. Effects of protected area downsizing on habitat fragmentation in Yosemite National Park (USA), 1864–2014. Ecology and Society 21(3):22. https://doi.org/10.5751/ES-08679-210322

Golden Kroner, R. E., S. Qin, C. N. Cook, R. Krithivasan, S. M. Pack, O. D. Bonilla, K. A. Cort-Kanisally, B. Coutinho, M. Feng, M. I. Martínez Garcia, Y. He, C. J. Kennedy, C. Lebreton, J. C. Ledezma, T. E. Lovejoy, D. A. Luther, Y. Parmanand, C. A. Ruíz-Aguldelo, E. Yerena, V. Morón Zambrano, and M. B. Mascia. 2019. The uncertain future of protected lands and waters. Science 364(6443) 881–886. https://doi.org/10.1126/science.aau5525

Gray, C. L., S. L. L. Hill, T. Newbold, L. N. Hudson, L. Börger, S. Contu, A. J. Hoskins, S. Ferrier, A. Purvis, and J. P. W. Scharlemann. 2016. Local biodiversity is higher inside than outside terrestrial protected areas worldwide. Nature Communications 7(1):12306. https://doi.org/10.1038/ncomms12306

Herrnstadt, E., and E. Muehlegger. 2014. Weather, salience of climate change and congressional voting. Journal of Environmental Economics and Management 68(3):435–448. https://doi.org/10.1016/j.jeem.2014.08.002

Kammerlander, A., and G. G. Schulze. 2021. Political-economic correlates of environmental policy. Environmental Research Letters 16(2):024047. https://doi.org/10.1088/1748-9326/abdc89

Keles, D., P. Delacote, A. Pfaff, S. Qin, and M. B. Mascia. 2020. What drives the erasure of protected areas? Evidence from across the Brazilian Amazon. Ecological Economics 176:106733. https://doi.org/10.1016/j.ecolecon.2020.106733

Keles, D., A. Pfaff, and M. B. Mascia. 2023. Does the selective erasure of protected areas raise deforestation in the Brazilian Amazon? Journal of the Association of Environmental and Resource Economists 10(4):1121–1147. https://doi.org/10.1086/723543

League of Conservation Voters (LCV). 2021. National environmental scorecard. 2000–2019. League of Conservation Voters, Washington, D.C., USA. https://scorecard.lcv.org/

Lewis, J. B., B. DeVine, L. Pitcher, and K. C. Martis. 2013. Digital boundary definitions of United States congressional districts, 1789-2017. >https://cdmaps.polisci.ucla.edu

List, J. A., and D. M. Sturm. 2006. How elections matter: theory and evidence from environmental policy. Quarterly Journal of Economics 121(4):1249–1281.

Mascia, M. B., and S. Pailler. 2011. Protected area downgrading, downsizing, and degazettement (PADDD) and its conservation implications. Conservation Letters 4(1):9–20. https://doi.org/10.1111/j.1755-263X.2010.00147.x

McAlexander, R. J., and J. Urpelainen. 2020. Elections and policy responsiveness: evidence from environmental voting in the US congress. Review of Policy Research 37(1):39–63. https://doi.org/10.1111/ropr.12368

Milman, O. 2019. Political polarisation over climate crisis has surged under Trump. 11 October, Guardian. https://www.theguardian.com/environment/2019/oct/11/political-polarisation-climate-crisis-trump

Mittermeier, R. A., and C. G. Mittermeier. 1997. Megadiversity: Earth’s biologically wealthiest nations. Cemex.

Naughton-Treves, L., and M. B. Holland. 2019. Losing ground in protected areas? Science 364(6443):832–833. https://doi.org/10.1126/science.aax6392

National Park Service. 2016. Gathering of certain plants or plant parts by federally recognized Indian Tribes for traditional purposes. Federal Register. Vol. 81(133):45024. Tuesday, July 12, 2016. https://www.federalregister.gov/documents/2016/07/12/2016-16434/gathering-of-certain-plants-or-plant-parts-by-federally-recognized-indian-tribes-for-traditional

Olsson, E., R. Albrecht, and R. E. Golden Kroner. 2021. PADDDtracker data release Version 2.1: technical notes. Conservation International, Arlington, Virginia, USA. https://data.subak.org/dataset/padddtracker-data-release-version-2-1

Pacca, L., D. Curzi, G. Rausser, and A. Olper. 2021. The role of party affiliation, lobbying, and electoral incentives in decentralized US state support of the environment. Journal of the Association of Environmental and Resource Economists 8(3):617–653. https://doi.org/10.1086/711583

Pachauri, R. K. , M. R. Allen, V. R. Barros, J. Broome, W. Cramer, R. Christ, J. A. Church, L. Clarke, Q. Dahe, P. Dasgupta, N. K. Dubash, O. Edenhofer, I. Elgizouli, C. B. Field, P. Forster, P. Friedlingstein, J. Fuglestvedt, L. Gomez-Echeverri, S. Hallegatte, G. Hegerl, M. Howden, K. Jiang, B. Jimenez Cisneroz, V. Kattsov, H. Lee, K. J. Mach, J. Marotzke, M. D. Mastrandrea, L. Meyer, J. Minx, Y. Mulugetta, K. O’Brien, M. Oppenheimer, J. J. Pereira, R. Pichs-Madruga, G. K. Plattner, H. O. Pörtner, S. B. Power, B. Preston, N. H. Ravindranath, A. Reisinger, K. Riahi, M. Rusticucci, R. Scholes, K. Seyboth, Y. Sokona, R. Stavins, T. F. Stocker, P. Tschakert, D. van Vuuren, and J. P. van Ypserle. 2014. Climate change 2014: synthesis report. Contribution of Working Groups I, II and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. IPCC, Geneva, Switzerland. https://epic.awi.de/id/eprint/37530/1/IPCC_AR5_SYR_Final.pdf

Pack, S. M., M. N. Ferreira, R. Krithivasan, J. Murrow, E. Bernard, and M. B. Mascia. 2016. Protected area downgrading, downsizing, and degazettement (PADDD) in the Amazon. Biological Conservation 197:32–39. https://doi.org/10.1016/j.biocon.2016.02.004

Pailler, S. 2018. Re-election incentives and deforestation cycles in the Brazilian Amazon. Journal of Environmental Economics and Management 88:345–365. https://doi.org/10.1016/j.jeem.2018.01.008

Peters, R., W. J. Ripple, C. Wolf, M. Moskwik, G. Carreón-Arroyo, G. Ceballos, A. Córdova, R. Dirzo, P. R. Ehrlich, A. D. Flesch, R. List, T. E. Lovejoy, R. F. Noss, J. Pacheco, J. K. Sarukhán, M. E. Soulé, E. O. Wilson, and J. R. B. Miller. 2018. Nature divided, scientists united: US-Mexico border wall threatens biodiversity and binational conservation. BioScience 68(10):740–743. https://doi.org/10.1093/biosci/biy063

Pfaff, A., F. Santiago-Ávila, and L. Joppa. 2017. Evolving protected-area impacts in Mexico: political shifts as suggested by impact evaluations. Forests 8(1):17. https://doi.org/10.3390/f8010017

Richards, R. 2018. Measuring conservation progress in North America. Center for American Progress, Washington, D.C., USA. https://www.americanprogress.org/article/measuring-conservation-progress-north-america/

Robalino, J., A. Pfaff, and L. Villalobos. 2017. Heterogeneous local spillovers from protected areas in Costa Rica. Journal of the Association of Environmental and Resource Economists 4(3):795–820. https://doi.org/10.1086/692089

Ruggiero, P. G., A. Pfaff, E. Nichols, M. Rosa, and J. P. Metzger. 2021. Election cycles affect deforestation within Brazil’s Atlantic Forest. Conservation Letters 14(5):e12818. https://doi.org/10.1111/conl.12818

Russell, N. J. 2006. An introduction to the Overton window of political possibilities. Mackinac Center for Public Policy, Midland, Michigan, USA. https://www.mackinac.org/7504

Schulze, K. 2021. Policy characteristics, electoral cycles, and the partisan politics of climate change. Global Environmental Politics 21(2):44–72. https://doi.org/10.1162/glep_a_00593

Shah, P., K. Baylis, J. Busch, and J. Engelmann. 2021. What determines the effectiveness of national protected area networks? Environmental Research Letters 16(7):074017. https://doi.org/10.1088/1748-9326/ac05ed

Symes, W. S., M. Rao, M. B. Mascia, and L. R. Carrasco. 2016. Why do we lose protected areas? Factors influencing protected area downgrading, downsizing and degazettement in the tropics and subtropics. Global Change Biology 22(2):656–665. https://doi.org/10.1111/gcb.13089

Tesfaw, A. T., A. Pfaff, R. E. Golden Kroner, S. Qin, R. Medeiros, and M. B. Mascia. 2018. Land-use and land-cover change shape the sustainability and impacts of protected areas. Proceedings of the National Academy of Sciences 115(9):2084–2089. https://doi.org/10.1073/pnas.1716462115

U.S. Bureau of Economic Analysis. 2020. GDP by county, metro, and other areas. BEA, Washington, D.C., USA. https://www.bea.gov/data/gdp/gdp-county-metro-and-other-areas

U.S. Department of the Interior. 2021. Conserving and restoring America the beautiful. U.S. Department of the Interior, Washington, D.C., USA. https://www.doi.gov/sites/default/files/report-conserving-and-restoring-america-the-beautiful-2021.pdf

U.S. Geological Survey (USGS) Gap Analysis Project (GAP). 2018. Protected areas database of the United States (PAD-US): U.S. Geological Survey data release. U.S. Geological Survey, Reston, Virginia, USA. https://doi.org/10.5066/P955KPLE