Sage Journals: Discover world-class research

Abstract

Since the mid-1980s, the US Border Patrol has positioned itself both as an enforcement agency and a provider of life-saving humanitarian assistance to migrants in distress. Over the past few decades — as the agency ratcheted up border enforcement — the number of migrant deaths along the US–Mexico border simultaneously increased. In response, the Border Patrol launched various humanitarian programs and initiatives, including deploying “rescue beacons” throughout the borderlands. These beacons are small towers that allow migrants to press a button and call for help. While the Border Patrol has championed these beacons as saving lives, there is no systematic evaluation of the relationship between rescue beacons and migrant deaths. This paper aims to better understand the Border Patrol's humanitarian activities and their effects by exploring the relationship between rescue beacons and migrant deaths in Pima County, Arizona from 2001 to 2022. Although the Border Patrol links rescue beacons to saving lives, this quantitative study finds little systematic evidence that the beacons reduce the discovery of recovered migrant remains in their vicinity.

Keywords

U.S.-Mexico border immigration enforcement migrant deaths Border Patrol rescue beacons

Introduction

During the past three decades, the Border Patrol has documented more than 10,000 migrant deaths along the US–Mexico border (Peterson 2024; U.S. Border Patrol 2024a, 2024b). These migrants died from various causes, including heat exposure and dehydration, drowning, and highway accidents. Migrant deaths are not a new phenomenon — dating back more than 100 years — but recent numbers are the highest on record.

In response to the deaths, the Border Patrol has engaged in targeted humanitarian assistance to migrants in distress. In the 1980s, it focused on training its agents with life-saving medical and rescue techniques (Hall 1987). However, by the late 1990s, the Border Patrol had moved into additional activities, including expanding its search and rescue efforts border-wide through the Border Patrol Search Trauma and Rescue (BORSTAR) unit, and constructing rescue beacons in remote border areas, which migrants can activate to receive assistance. In recent years, the Border Patrol expanded these efforts through its Missing Migrants Program, and has continued deploying rescue beacons and geolocated emergency placards along the border.

This article focuses on the Border Patrol's use of rescue beacons. Rescue beacons are tower-like structures with a button that migrants can push to request help. As of August 2023, the agency had deployed 174 beacons in remote areas along the border, and has routinely promoted the beacons as “saving lives” and serving as migrants’ “only chance of survival” in the desert terrain (U.S. Customs and Border Protection 2021a, 2021b, 2023, 12). However, the Border Patrol has never formally evaluated the relationship between these beacons and migrant deaths. Instead, the agency tracks the number of beacon activations and associated rescues (U.S. Customs and Border Protection 2023, 12). While these rescues are important, as some of these individuals may have died without assistance, this metric alone does not capture the full relationship between rescue beacons and migrant deaths.

This article serves as an exploratory quantitative study to assess the relationship between rescue beacons and migrant deaths. It focuses on 26 rescue beacons in Pima County, Arizona. This area was chosen due to its status as the deadliest border crossing area for migrants over the last two decades and the availability of rescue beacon data across the county. The article also uses geolocated migrant death data from the Pima County Medical Examiner's Office.

The article finds that despite the Border Patrol's humanitarian objectives in deploying rescue beacons, there is little statistical evidence that they reduce the discovery of recovered migrant remains in their vicinity. In some areas, the deployment of a beacon was even correlated with an increase in the recovery of migrant remains in the surrounding area. These changes may be due to shifting migration routes, as migrants and smugglers associate the beacons with surveillance measures and seek to find new routes around them. Alternatively, they may be related to migrants moving their routes closer to the beacons and increasing migration in the surrounding areas, or authorities’ increased foot traffic around the beacons, leading to more remains discoveries. While the Border Patrol's data on migrant rescues show that hundreds of people have successfully used rescue beacons, the article's findings suggest that their efficacy is more nuanced.

Migrant Deaths Along the Arizona–Mexico Border

For more than 100 years, migrants have died along the US–Mexico border. These dynamics date back to the 1880s, when the US Congress passed its first restrictive immigration policies and unauthorized migrants began crossing the US–Mexico border. While most of these early migrants successfully arrived at their final destinations, newspapers reported cases of migrants dying in California's deserts, drowning in the Rio Grande, and suffocating in train boxcars as far back as the 1890s (Los Angeles Herald 1890; El Paso Herald 1911; Yorkville Enquirer 1912).

Over time, border enforcement efforts made clandestine crossings more challenging. This culminated in 1994, when the Border Patrol launched its “Prevention Through Deterrence” (PTD) strategy, which concentrated agents in urban areas and explicitly aimed to push migrants into the border's most remote zones (U.S. Border Patrol 1994). As a result of this strategy, scholars began documenting an increase in the number of people traversing rougher, rural terrain and dying from environmental factors (e.g., heat exposure) (Eschbach et al. 1999; Cornelius 2001). These changes coincided with increases in the total number of migrant deaths (U.S. Government Accountability Office 2006).

The changing enforcement dynamics shaped migration routes and migrant deaths’ geospatial distribution across Arizona. First, as the Border Patrol's PTD strategy rerouted migrants away from the popular El Paso and San Diego crossing locations, more migrants began arriving to cities along the Arizona border, which scholars call the “funnel effect” (Rubio-Goldsmith et al. 2006; Soto and Martínez 2018; Martínez et al. 2021). Second, as more migrants began crossing in Arizona, the Border Patrol increased its enforcement efforts in the state's border cities, such as Nogales and Douglas, and manned round-the-clock highway checkpoints. These efforts created “secondary funnel effects” and “tertiary funnel effects” that pushed migrants into the surrounding desert (Chambers et al. 2021, 2023; Martínez et al. 2021; Chambers, Boyce and Martínez 2022). Finally, since 2014, continually expanding Border Patrol enforcement throughout the borderlands has created a “localized funnel effect” that shifts migrants’ routes into the most remote and treacherous areas in the state, such as the Tohono O’odham Nation and western Pima County (Martínez et al. 2021; Chambers, Boyce and Martínez 2022; Chambers et al. 2023).

In recent years, the number of migrant deaths along the US–Mexico border has reached record highs. From fiscal years 1998 to 2020, the Border Patrol reported an average of 350 migrant deaths yearly (U.S. Border Patrol 2024b). However, during fiscal years 2021 to 2023, these numbers more than doubled to 717 average deaths, with nearly 20 percent occurring in Arizona's Tucson Sector (Peterson 2024). The higher numbers are likely due in large part to the sharp increase in the number of migrants crossing the border during these years. Yet, these figures also make clear that migrant deaths remain a humanitarian issue and policy challenge.

The Border Patrol's Response to Migrant Deaths and Rescue Beacons

As the number of migrant deaths has increased in recent decades, border agencies have at times responded by implementing activities focused on risk reduction. These activities include developing targeted humanitarian technologies, rescuing migrants in distress, and counting deceased migrants (Williams 2013, 2016; Aas and Gundhus 2015; Pallister-Wilkins 2015; Perkowski 2018).

The US Border Patrol has been aware of migrant deaths along its border with Mexico since the 1970s but announced its first humanitarian efforts — the Yuma Sector's Desert Area Rescue Team and the Tucson Sector's Star Team — a decade later in 1986 (Los Angeles Times 1979; Ralles 1986; Hall 1987). These two teams of specialized agents engaged in rescues, emergency first aid, and migrant death documentation. However, they were hamstrung by their small number of trained agents compared to the vast territory in which they operated.

In 1998, the Border Patrol launched its first border-wide initiative for addressing migrant deaths: the Border Safety Initiative. This initiative came as the agency implemented its PTD strategy and more migrants began dying from exposure to the elements (Eschbach et al. 1999; Cornelius 2001). The Border Safety Initiative was designed to improve migrant safety, and included deploying agents and surveillance technology to dangerous areas, placing warning signs along routes, engaging in public information campaigns, and launching the agency's BORSTAR units (U.S. Department of Justice 1998). Notably, it also called for placing the first rescue beacons along the US–Mexico border (U.S. Department of Justice 1998).

This article focuses on these rescue beacons, which are roughly 35 feet tall, self-contained units that rely on solar power. Each beacon is equipped with a blue strobe light at the top, designed to be visible for multiple miles. A sign instructs migrants in English, Spanish, and through visual depictions to press a red button to call for help.

Over the past 20 years, the Border Patrol has steadily added rescue beacons along the US–Mexico border. In March 2002, the Border Patrol placed the first six beacons in the Yuma desert (Klug 2002). By 2005, it had constructed at least eight within its Tucson Sector, which includes Pima County (Winograd 2005). In 2019, the Missing Persons and Unidentified Remains Act required that the Border Patrol “purchase, deploy, and maintain” up to 170 rescue beacons in border locations (U.S. Congress 2019). As of August 2023, the Border Patrol had deployed 174 rescue beacons, with approximately 20 percent in Arizona's Tucson Sector (U.S. Customs and Border Protection 2023).

To determine where to place a rescue beacon, the Border Patrol considers several factors. As of June 2021, the agency used a model with weighted operational and environmental variables (U.S. Government Accountability Office 2022, 13). These include traffic patterns (a 20% weight), discovered migrant deaths (20%), land cover types (20%), slope (15%), previous rescues (10%), direction from discovered remains (5%), elevation (5%), and proximity to roads (5%). In May 2018, the Cabeza Prieta Wildlife Refuge's website listed that beacons were deployed to areas with high levels of migration and migrant deaths (Plath 2020).

The Border Patrol has consistently noted that its beacons are “saving lives” (U.S. Customs and Border Protection 2021b). For example, the agency notes that in fiscal year 2022, it conducted “214 successful rescues associated with activated rescue beacons” (U.S. Customs and Border Protection 2023, 12). This reported effectiveness has also made its way into courtroom arguments. In 2019, a US federal prosecutor argued that civil society groups should leave all humanitarian aid to the Border Patrol, and pointed to the deployed rescue beacons as saving migrants’ lives (Hennessy-Fiske 2019).

However, civil society organizations and scholars have cast doubt on this effectiveness. In a February 2021 report, a civil society organization criticized the low number of beacons throughout the Arizona desert and noted that “there are numerous deaths in close proximity to Border Patrol rescue beacons” (No More Deaths and La Coalición de Derechos Humanos 2021, 37). They also criticized the beacons’ placement and wrote that “most rescue beacons are not placed in the most remote and deadly borderlands migration trails, but on roadways where they can be easily accessed by line agents and serviced for repairs” (No More Deaths and La Coalición de Derechos Humanos 2021, 37). Researchers have also suggested that the beacons may push migrants into even more dangerous paths “if they are perceived not as a point of aid but rather as a threat of deportation” (Plath 2020).

Over the past three decades, scholars have repeatedly linked the Border Patrol's enforcement efforts — such as more agents, highway checkpoints, and border fencing — to shifting migration patterns (Bean et al. 1994; DePalma 1995; Rubio-Goldsmith et al. 2006; Martínez et al. 2018; Massey, Pren and Durand 2016; Chambers et al. 2021; Bansak et al. 2024). Scholars also documented how migrants sought to avoid Border Patrol surveillance towers by moving “into increasingly difficult and mountainous areas that lie adjacent” to the towers (Chambers et al. 2021, 460). Notably, the surveillance towers look similar to rescue beacons, as both are tall metal structures with cameras and lights. Yet, to date, scholars have not evaluated whether the Border Patrol's rescue beacons might generate similar effects.

Data

We combine several data sources to evaluate the relationship between the Border Patrol's rescue beacons and migrant deaths in Pima County. A primary challenge in studying rescue beacons is the dearth of information on their locations and operation dates. Citing national security concerns, the Border Patrol does not release detailed information about its rescue beacons. We obtained location information for some beacons through Freedom of Information Act (FOIA) requests and scholars’ previous work.

To identify rescue beacons within Pima County, we combined and cross-referenced various data sources. In March 2024, the Border Patrol responded to an FOIA request with a list of the 34 active rescue beacons in its Tucson Sector, which includes Pima, Santa Cruz, Cochise, Maricopa, and Pinal counties (U.S. Customs and Border Protection 2024). The Border Patrol did not include GPS coordinates for the beacons. However, we determined that most of the active beacons were in Pima County by cross-referencing the list against a Border Patrol map from August 2023, which displayed rescue beacon locations along the border (U.S. Customs and Border Protection 2023, 23–24). In this map, the Border Patrol's rescue beacons are almost exclusively clustered in the Tucson Sector's western zone, which is made up of Pima County. Using this cross-referencing, we were able to use the data we received through the FOIA request to create an upper bound number of active rescue beacons in Pima County.

Notably, while the Border Patrol's March 2024 FOIA response did not include the rescue beacons’ precise locations, it did indicate the land-management entity for each beacon. This variable allowed us to tally beacon subtotals across different areas in Pima County, such as Organ Pipe National Monument, Cabeza Prieta National Wildlife Refuge, and the Tohono O’odham Nation's land. It also allowed us to determine that one of the Border Patrol's listed rescue beacons was located just north of Pima County — in the Barry M. Goldwater Airforce Range in Maricopa County — and to remove it from our analysis.

We then built our list of rescue beacon locations in Pima County. First, we used information from a January 2024 FOIA request to the US Department of the Interior that included the locations of rescue beacons in the Cabeza Prieta Wildlife Refuge (U.S. Department of the Interior 2024). We determined that eight of the Cabeza Prieta Wildlife Refuge's 13 rescue beacons were located in Pima County. This matched the Border Patrol's total for the Cabeza Prieta Wildlife Refuge in its March 2024 FOIA response. We then used a March 2022 FOIA response from the US Department of the Interior that provided the coordinates of nine rescue beacons in Organ Pipe National Monument, all of which were located in Pima County (U.S. Department of the Interior 2022). This number also matched the Border Patrol's total for Organ Pipe National Monument.

For the remaining beacons in the Tohono O’odham Nation's land and the Buenos Aires National Wildlife Refuge, we relied on a dataset that researcher Tara Plath compiled and made publicly available (Plath 2024).¹ Through her dataset, we added seven active beacons and two previously active beacons. In total, our dataset identified 24 active rescue beacons and two previously active beacons in Pima County (Figure 1 shows a map), accounting for 73 percent of the Border Patrol's active beacons in Pima County and 100 percent in western Pima County. The nine active rescue beacons not in our sample are located in the Tohono O’odham Nation's land and the Buenos Aires National Wildlife Refuge in eastern Pima County.

Figure 1.

Map of Dataset's Rescue Beacons in Pima County.

In addition to beacon locations, we required their placement timing. The Border Patrol does not publish the operation dates for its rescue beacons, so we used open satellite imagery — through Google Earth — to approximate them. However, satellite data is limited by the public images’ frequency (e.g., images of a given location can be separated by multiple years). To illustrate how we use satellite images to infer approximate operation dates, consider the example in Figure 2. Assume that each of the ovals and circles represents a satellite image of the same location taken on different dates. The shaded ovals represent images with no visible rescue beacons and the circles marked with an “X” denote images with a visible rescue beacon. In this example, we infer that this location did not have a rescue beacon through June 1, 2007. We also infer that this location did have a rescue beacon in operation from September 1, 2010 through January 1, 2019. Further, we infer that this location once again did not have a beacon from February 1, 2021 through July 7, 2023.² For our analysis, we exclude the dates (for each beacon) when we are uncertain about the beacon's placement and operation.

Figure 2.

Example: Inferring Approximate Operation Dates.

Information on recovered migrant deaths comes from the Pima County Office of the Medical Examiner (PCOME). We accessed the data through the civil society organization Humane Borders’ online portal, which publishes the PCOME's migrant death data each month through its online platform (Humane Borders 2024). We downloaded these data on November 16, 2023 and used them to identify recovered remains for the period 2001 to 2022.³ In total, the data cover 3,761 recovered migrant remains. For each case of recovered remains, the data include the GPS coordinates associated with the recovery location, the recovery date, and a qualitative assessment of the remains’ state of decomposition.

Notably, the PCOME's data, like all migrant death data, are incomplete and imperfect (Rubio-Goldsmith et al. 2006; Soto and Martínez 2018; Martínez et al. 2021). The true number of migrant deaths in a given area generally depends on the number of individuals passing through that area, and geographic, environmental, and enforcement conditions, among other factors. Moreover, for a migrant death to appear in official data, like the PCOME's data, the remains must be found. These factors may interact with each other in ways that affect measurement. For example, in more remote border areas, migrants may face a higher risk of death and a lower probability that remains are discovered.

Empirical Model and Results

To conduct our analysis, we organize the study area into 86 square grid cells with sides that measure 16 km.⁴ For a visual depiction of these grid cells, see Figure 3. For data completeness, we only consider cells that are fully contained within Pima County. We then create one observation for each cell and day over the sample period from 2001 to 2022. Next, we combine this grid-by-day data with the migrant remains data using the spatial coordinates and the date of recovery. Specifically, for each cell and day, we create a series of binary variables that denote the recovery of remains, or lack thereof, on that day and in that cell. We repeat this analysis for all cells and days during the sample period.

Figure 3.

Pima County Grid Cells and Rescue Beacons.

We next identify the 18 cells in Pima County that contain at least one rescue beacon. We refer to these grid cells as “beacon cells.” To each beacon cell, we match at most 24 (nonbeacon) cells within Pima County. Eight of these are the cells immediately adjacent to the beacon cell (neighbors), and the remaining 16 cells are those adjacent to these first eight neighboring cells. If a (nonbeacon) cell is close to two beacon cells, then that cell is matched to both beacon cells. Next, we incorporate the satellite data to measure when the beacon associated with each beacon cell was in operation.

After conducting this match for each of the 18 beacon cells, we collapse by month and stack the resulting data. We then characterize the relative location of each matched cell to its assigned beacon cell(s) using a set of binary variables. These binary variables measure the presence of a beacon cell (i.e., the cell itself contains one or more beacons) and the close proximity to the assigned beacon cell (i.e., the cell is immediately adjacent to the closest beacon cell).

We estimate linear regression models to characterize the relationship between rescue beacons and the recovery of migrant remains. All models include year-fixed effects to account for aggregate trends in migrant deaths and month of year-fixed effects to account for seasonality.⁵ We cluster standard errors by beacon cells. Our primary outcome variable is a binary variable indicating the recovery of migrant remains in cell i in month t. We first estimate a parsimonious specification that includes an indicator variable denoting a beacon cell, an indicator variable for months after which the beacon is operational, the interaction of these two variables, and the fixed effects.⁶

We distinguish between remains at differing stages of decomposition using the qualitative information available in the PCOME's data to create five dependent variables. These variables characterize the recovered remains as: (1) all remains regardless of decomposition state, (2) fully fleshed or decomposing remains, (3) fully fleshed remains, (4) decomposing remains, and (5) skeletal remains.⁷ The results are reported in Table 1. In four of the columns, the estimated coefficient on the beacon cell indicator is positive and statistically significant, which is consistent with the Border Patrol's siting formula that accounts for the past discovery of migrant remains. Further, in all columns, the positive and statistically significant estimated coefficient on the post beacon indicator suggests an increase in the likelihood of remains recovery over time.

Table 1.

Main Results with Full Sample of Pima County Beacons.

	(1) All	(2) Fleshed/decomposed	(3) Fully fleshed	(4) Decomposed	(5) Skeletal
Has beacon	0.0415** (0.0144)	0.0259** (0.0104)	0.0118 (0.00723)	0.0181** (0.00709)	0.0228*** (0.00761)
Postbeacon	0.0425** (0.0175)	0.0290** (0.0109)	0.00821** (0.00368)	0.0221** (0.00865)	0.0180* (0.0101)
Has beacon × Postbeacon	0.0674* (0.0346)	0.0285 (0.0194)	−0.00257 (0.00647)	0.0300* (0.0165)	0.0462* (0.0239)
N	56,120	56,120	56,120	56,120	56,120
Mean	0.074	0.046	0.019	0.031	0.029
Standard deviation	0.261	0.210	0.137	0.172	0.168

Note: Standard errors in parentheses.

*p < .10. **p < .05. ***p < .01.

Notably, the sign and significance of the interaction variable (i.e., the indicator for a cell likely having an operational beacon) varies across the dependent variables. The estimated interaction variable is statistically significant for three of the five dependent variables, indicating an increase in the likelihood that remains are recovered in the immediate area following a beacon's placement.

We explore robustness of the results in Table 1 by restricting our sample to western Pima County, where we are confident that our sample contains all active rescue beacons. Given our lack of complete coverage in eastern Pima County, we remove this area from our sample, which means removing beacons 13, 14, 15, 16, and 26 in Figure 1. We then re-estimate the model.⁸ The results are reported in Table 2. The results are qualitatively similar to our main results, with the estimated coefficients on the interaction variable now positive for all five models, and statistically different from zero in four models.

Table 2.

Main Results with Subsample of Beacons Located in Western Pima County.

	(1) All	(2) Fleshed/decomposed	(3) Fully fleshed	(4) Decomposed	(5) Skeletal
Has beacon	0.0385*** (0.0121) [0.00]	0.0217** (0.00844) [0.038]	0.00610 (0.00447) [0.218]	0.0154*** (0.00502) [0.022]	0.0242*** (0.00713) [0.00]
Postbeacon	0.0240*** (0.00611) [0.002]	0.0106*** (0.00228) [0.00]	0.00207*** (0.000505) [0.00]	0.00853*** (0.00242) [0.00]	0.0132** (0.00601) [0.054]
Has beacon × Postbeacon	0.0979** (0.0446) [0.058]	0.0443* (0.0223) [0.09]	0.00270 (0.00532) [0.662]	0.0428* (0.0201) [0.066]	0.0603* (0.0318) [0.082]
N	34,020	34,020	34,020	34,020	34,020

Note: Standard errors in parentheses.

*p < .10. **p < .05. ***p < .01.

Thus, restricting our sample to western Pima County, we find a more consistent relationship between a rescue beacon's placement and the recovery of migrant remains. Specifically, the results in Table 2 indicate that after a beacon's placement in western Pima County, there is an increase in the recovery of migrant remains in the immediate surrounding areas. The difference between these results and those in Table 1 suggests that the missing beacons in eastern Pima County may bias our main estimates.

Next, we explore the patterns across all of Pima County for recovered remains in the grid cells that neighbor beacon cells. We conduct this analysis by adding an indicator variable for cells that are immediately adjacent to beacon cell(s) — a neighbor — as well as an interaction between this variable and the post beacon indicator. The results are reported in Table 3. First, we find that the estimated coefficients in the first three rows are similar to our main results (i.e., Table 1). Second, in all five models, the estimated positive coefficient on the neighbor indicator suggests a higher likelihood of remains being discovered in cells that neighbor a beacon cell — compared to those that are located further away — prior to the beacon's placement. Third, the estimated coefficient on the neighbor postplacement interaction term is statistically indistinguishable from zero. Thus, we find no changes in the likelihood of remains being discovered in neighboring cells following the beacon's placement in a nearby cell. This suggests that if smugglers and migrants are shifting their routes to circumvent rescue beacons, they are likely making only small adjustments within the 16 × 16 km² beacon cells.

Table 3.

Closest Neighbor Results with Full Sample of Pima County Beacons.

	(1) All	(2) Fleshed/decomposed	(3) Fully fleshed	(4) Decomposed	(5) Skeletal
Has beacon	0.0506*** (0.0164)	0.0329** (0.0118)	0.0156* (0.00830)	0.0219** (0.00765)	0.0257*** (0.00811)
Postbeacon	0.0382** (0.0171)	0.0270** (0.0106)	0.00966*** (0.00320)	0.0190** (0.00868)	0.0148 (0.0103)
Has beacon × Postbeacon	0.0701* (0.0348)	0.0293 (0.0196)	−0.00436 (0.00703)	0.0323* (0.0164)	0.0486* (0.0239)
Neighbor	0.0314*** (0.0104)	0.0239*** (0.00721)	0.0130** (0.00480)	0.0133*** (0.00432)	0.00988** (0.00426)
Neighbor × Postbeacon	0.00716 (0.0116)	0.00159 (0.00609)	−0.00645 (0.00423)	0.00692 (0.00491)	0.00733 (0.00673)
N	56,120	56,120	56,120	56,120	56,120

Note: Standard errors in parentheses.

*p < .10. **p < .05. ***p < .01.

As a final exercise, we estimate a separate linear regression for each beacon cell to allow for variation in the relationship between a rescue beacon's placement and the discovery of remains across different locations in Pima County. The dependent variable indicates the discovery of any migrant remains, which is consistent with the columns labeled “All” in Tables 1 to 3. The results are reported in Tables 4 and 5. Of the 18 beacon cell-specific models, the interaction coefficient is positive and statistically significant in seven models, negative and statistically significant in four models, and statistically indistinguishable from zero in seven models. Thus, we find evidence of variation in the relationship between a beacon's placement and migrant remains recoveries across beacons located in different areas of Pima County. However, the pattern of effects is unclear, with positive, negative, and null effects in both western and eastern Pima County.

Table 4.

Beacon Cell-Specific Results.

	(1)	(2)	(3, 5, 9)	(4, 21)	(6, 7)	(8)	(10)	(11)	(12)
Has beacon	0.0264** (0.0123)	−0.0230*** (0.00613)	0.0374* (0.0205)	−0.0125 (0.0336)	0.0185 (0.0122)	0.00589 (0.00965)	0.0707*** (0.0249)	0.0790*** (0.0290)	0.0362 (0.0250)
Postbeacon	0.00521** (0.00251)	0.000846 (0.0160)	−0.00876 (0.0102)	0.00208 (0.0216)	−0.00700 (0.00649)	−0.00874 (0.00543)	−0.00524 (0.0123)	−0.00620 (0.0183)	0.0478** (0.0223)
Has beacon × Postbeacon	−0.0332** (0.0133)	0.185*** (0.0656)	0.531*** (0.0696)	0.227*** (0.0576)	0.112 (0.0713)	0.114 (0.0696)	−0.0204 (0.0376)	−0.0737* (0.0383)	−0.0152 (0.0373)
N	2,376	1,071	2,210	1,331	3,312	2,691	2,856	3,030	4,016

Note: Standard errors in parentheses.

*p < .10. **p < .05. ***p < .01.

Table 5.

Beacon Cell-Specific Results.

	(13)	(14)	(15)	(16)	(17)	(18, 19, 20, 22, 23)	(24)	(25)	(26)
Has beacon	−0.0365* (0.0216)	0.117*** (0.0318)	0.327*** (0.0548)	0.0466 (0.0299)	0.0170 (0.0172)	0.0456* (0.0274)	0.150*** (0.0324)	0.0298 (0.0249)	0.0111 (0.0396)
Postbeacon	0.0835** (0.0382)	0.0714*** (0.0250)	0.0713*** (0.0245)	−0.0418 (0.0906)	0.0209 (0.0142)	0.0331 (0.0336)	0.0458* (0.0236)	0.0457* (0.0266)	0.103*** (0.0233)
Has beacon × Postbeacon	−0.0440 (0.0269)	−0.0231 (0.0615)	−0.164** (0.0657)	0.247*** (0.0560)	0.186*** (0.0477)	0.239*** (0.0576)	−0.103** (0.0471)	0.0730* (0.0423)	−0.0121 (0.0472)
N	4,338	4,199	4,788	3,825	2,662	1,624	3,630	3,211	4,950

Note: Standard errors in parentheses.

*p < .10. **p < .05. ***p < .01.

Conclusion

Since the mid-1980s, the US Border Patrol has positioned itself as both an enforcer of US immigration laws and as a provider of humanitarian assistance to migrants in distress. In line with this latter mission, for the past two decades, the Border Patrol has constructed rescue beacons along the US–Mexico border. In recent years, US officials have pointed to the rescue beacons as saving migrants’ lives (Hennessy-Fiske 2019; U.S. Customs and Border Protection 2021a). However, to date, there has been no rigorous evaluation of the relationship between the Border Patrol's rescue beacons and migrant deaths. This paper provides an exploratory analysis of this relationship. We find little evidence that the beacons reduce the discovery of recovered migrant remains in their vicinity.

Overall, the results show a nuanced relationship between rescue beacons and migrant deaths. First, a beacon's placement is weakly associated with an increase in the recovery of migrant remains in the immediate surrounding area. In western Pima County, where we are more confident in our coverage of active beacons, the positive association is more consistent. Second, after a beacon is installed, there is no evidence that remains are more or less likely to be discovered in neighboring cells. Third, our beacon-level analysis suggests that the relationship between beacons and the recovery of migrant remains varies across locations within Pima County.

There are multiple explanations for this exploratory analysis's pattern of results. First, migrants may adjust their routes to circumvent beacons, and these new routes may be more remote and deadly. Previous studies documented how migrants changed their migration routes in response to the Border Patrol's construction of border fencing and surveillance towers throughout the borderlands (Chambers et al. 2021; Bansak et al. 2024). However, if smugglers and migrants are shifting their routes to circumvent rescue beacons, they must only make relatively small adjustments — given that we find no evidence of a relationship between a beacon's placement and the recovery of remains in neighboring cells. Second and alternatively, migrants may adjust their routes to move closer to beacons, increasing migrant flows in the area and the number of remains recovered even with no change in the rate of migrant deaths. Finally, a beacon's placement may increase authorities’ foot traffic in the surrounding areas (e.g., by Border Patrol agents), which then increases the likelihood that migrant remains are discovered.

Our study has several limitations. First, we do not have access to geolocated migrant apprehension data, which would allow us to approximate migration flows across and within cells over time. This limitation prevents us from studying how rescue beacons affect the rate of migrant deaths and thereby parsing the potential mechanisms that underlie our results. Second, we do not observe the true number of migrants who die in the study area, and, instead, rely on data that documents only recovered migrant remains. Third, we do not have complete information on all rescue beacon locations in Pima County. Fourth, we lack precise information on rescue beacons’ exact operating dates. Finally, our exploratory analysis is limited to a specific segment of the US–Mexico border. Thus, our results may not extend to other areas of the border.

In conclusion, we find no consistent evidence that rescue beacons reduce the likelihood of recovered remains in their vicinity. However, Border Patrol data on associated rescues highlights how hundreds of individuals have used beacons to receive potentially life-saving help and medical assistance. Overall, there is a need for additional scholarship to continue exploring the relationship between rescue beacons and migrant deaths, and how border agencies’ humanitarian activities affect migrant safety.

Supplemental Material

sj-docx-1-mrx-10.1177_01979183251390572 - Supplemental material for Rescue Beacons and Migrant Deaths in Southern Arizona

Supplemental material, sj-docx-1-mrx-10.1177_01979183251390572 for Rescue Beacons and Migrant Deaths in Southern Arizona by Stephanie Leutert, Ludovica Gazze, and Mary Evans in International Migration Review

Footnotes

Acknowledgements

We thank Daniel Martínez and Samuel Chambers for helpful suggestions. All errors and omissions remain our own.

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The authors received no financial support for the research,authorship,and/or publication of this article.

ORCID iDs

Stephanie Leutert

Ludovica Gazze

Mary Evans

Supplemental Material

Supplemental material for this article is available online.

References

Aas

K. F.

Gundhus

H. O. I.

. 2015. “Policing Humanitarian Borderlands: Frontex, Human Rights and the Precariousness of Life.” The British Journal of Criminology 55 (1): 1–18. https://doi.org/10.1093/bjc/azu086.

Norton

E. C.

. 2003. “Interaction Terms in Logit and Probit Models.” Economics Letters 80 (1): 123–9. https://doi.org/10.1016/S0165-1765(03)00032-6.

Bansak

Hall

Coon

Dieringer

. 2024. “Border Walls and Death on the US–Mexico Border: A Hot-Spot Analysis of the Impact of the Secure Fences Act in Arizona.” (Working Paper). The Center for Growth and Opportunity at Utah State University.

Bean

F. D.

Chanove

Cushing

R. G.

De La Garza

Freeman

G. P.

Haynes

C. W.

Spener

. 1994. Illegal Mexican Migration and the United States/Mexico Border: The Effects of Operation Hold the Line on El Paso/Juarez. Washington, DC: US Commission on Immigration Reform.

Cameron

A. C.

Gelbach

J. B.

Miller

D. L.

. 2008. “Bootstrap-Based Improvements for Inference with Clustered Errors.” The Review of Economics and Statistics 90 (3): 414–27. https://doi.org/10.1162/rest.90.3.414.

Chambers

S. N.

Boyce

G. A.

Launius

Dinsmore

. 2021. “Mortality, Surveillance and the Tertiary Funnel Effect on the US–Mexico Border: A Geospatial Modeling of the Geography of Deterrence.” Journal of Borderlands Studies 36 (3): 443–68. https://doi.org/10.1080/08865655.2019.1570861.

Chambers

S. N.

Boyce

G. A.

Martínez

D. E.

. 2022. “Climate Impact or Policy Choice? The Spatiotemporality of Thermoregulation and Border Crosser Mortality in Southern Arizona.” The Geographical Journal 188 (3): 401–14. https://doi.org/10.1111/geoj.12443.

Chambers

S. N.

Boyce

G. A.

Martínez

D. E.

Bongers

C. C. W. G.

Keith

. 2023. “The Contribution of Physical Exertion to Heat-Related Illness and Death in the Arizona Borderlands.” Spatial and Spatio-Temporal Epidemiology 46: Article 100590. https://doi.org/10.1016/j.sste.2023.100590.

Cornelius

W. A

. 2001. “Death at the Border: Efficacy and Unintended Consequences of US Immigration Control Policy.” Population and Development Review 27 (4): 661–85. https://doi.org/10.1111/j.1728-4457.2001.00661.x.

10.

DePalma

1995. “Countries Plan Joint Border Effort.” The San Bernardino County Sun 3.

11.

El Paso Herald. 1911. “Another Chinaman is Found in the River,” 11.

12.

Eschbach

Hagan

Rodriguez

Hernandez-Leon

Bailey

. 1999. “Death at the Border.” International Migration Review 33 (2): 430–54. https://doi.org/10.1177/019791839903300206.

13.

Hall

1987. “US Law, Mexican Alert, Cut Desert’s Death Toll.” Arizona Republic 21.

14.

Hennessy-Fiske

2019. “Should Activists Aid Migrants in the Desert, or Leave Their Fate to the Border Patrol?” Los Angeles Times. https://www.latimes.com/nation/la-na-arizona-migrant-rescue-trial-20190308-story.html

15.

Humane Borders. 2024. “Arizona OpenGIS Initiative for Deceased Migrants.” https://www.humaneborders.org/migrant-death-mapping

16.

Klug

2002. “Beacons Installed to Aid Migrants in Distress.” Associated Press in the Salt Lake Tribune 19.

17.

Los Angeles Herald. 1890. “Died in the Desert: The Fate of a Chinese Boy Who Tried to Cross.” Los Angeles Herald 2.

18.

Los Angeles Times. 1979. “U.S. Wonders How Many Mexicans Die at the Border?” Los Angeles Times Service 54.

19.

Martínez

D. E.

Reineke

R. C.

Boyce

Chambers

S. N.

Launius

Anderson

B. E.

Hess

G. L.

, J. M., et al. 2021. “Migrant Deaths in Southern Arizona: Recovered Undocumented Border Crosser Remains Investigated by the Pima County Office of the Mexican Examiner, 1990–2020.” Report. Binational Migration Institute, University of Arizona.

20.

Martínez

D. E.

Reineke

R. C.

Rubio-Goldsmith

Parks

B. O.

. 2018. “Structural Violence and Migrant Deaths in Southern Arizona: Data from the Pima County Office of the Medical Examiner, 1990–2013.” Journal on Migration and Human Security 2 (4): 257–86. https://doi.org/10.1177/233150241400200401.

21.

Massey

D. S.

Pren

K. A.

Durand

. 2016. “Why Border Enforcement Backfired.” American Journal of Sociology 121 (5): 1557–600. https://doi.org/10.1086/684200.

22.

No More Deaths and La Coalición de Derechos Humanos. 2021. “Left to Die: Border Patrol, Search and Rescue, & the Crisis of Disappearance.” http://www.thedisappearedreport.org/uploads/8/3/5/1/83515082/left_to_die_-_english.pdf

23.

Pallister-Wilkins

2015. “The Humanitarian Politics of European Border Policing: Frontex and Border Police in Evros.” International Political Sociology 9: 53–69. https://doi.org/10.1111/ips.12076.

24.

Perkowski

. 2018. “Frontex and the Convergence of Humanitarianism, Human Rights and Security.” Security Dialogue 49 (6): 457–75. https://doi.org/10.1177/0967010618796670.

25.

Peterson

2024. “FOIA of CBP BSITS Data.” El Paso Migrant Death Database. https://www.elpasomigrantdeathdatabase.org/index.php/foia/

26.

Pima County Office of the Medical Examiner. 2024. “Annual Report 2024.” https://content.civicplus.com/api/assets/31e65356-defa-4d88–86c9-cce40ae253d2?cache=1800.

27.

Plath

2020. “An Elusive Viewshed: An Investigation of United States’ Border Patrol Rescue Beacons in Arizona’s Western Desert.” Plot(s) Journal of Design Studies 7 (2): 26–42.

28.

Plath

2024. “AZ Beacon Map.” https://azbeaconmap.org/about/

29.

Ralles

1986. “Border Patrol’s Tucson District Creates Desert-Rescue Team.” Arizona Republic 11.

30.

Rubio-Goldsmith

McCormick

Martínez

Duarte

. 2006. The ‘Funnel Effect’ Recovered Bodies of Unauthorized Migrants Processed by the Pima County Office of the Medical Examiner, 1990–2005. Tucson, Arizona, United States: Binational Migration Institute, University of Arizona.

31.

Soto

Martínez

D. E.

. 2018. “The Geography of Migrant Death: Implications for Policy and Forensic Science.” In Sociopolitics of Migrant Death and Repatriation: Perspectives from Forensic Science, edited by Latham

O’Daniel

, 67–82. Cham, Switzerland: Springer International Publishing AG.

32.

U.S. Border Patrol. 1994. Border Patrol Strategic Plan: 1994 and Beyond. Washington, DC, United States: Homeland Security Digital Library.

33.

U.S. Border Patrol. 2024a. “Border Rescues and Mortality Data.” https://www.cbp.gov/newsroom/stats/border-rescues-and-mortality-data

34.

U.S. Border Patrol. 2024b. “Southwest Border Deaths by Fiscal Year.” https://www.cbp.gov/sites/default/files/assets/documents/2021-Aug/U.S.%20Border%20Patrol%20Fiscal%20Year%20Southwest%20Border%20Sector%20Deaths%20%28FY%201998%20-%20FY%202020%29%20%28508%29.pdf

35.

U.S. Congress. 2019. The Missing Persons and Unidentified Remains Act. Pub. L. No. 116–277. Washington, DC, United States: U.S. Federal Legislation.

36.

U.S. Customs and Border Protection. 2021a. “RGV Agents Deploy Life Saving Beacons.” https://www.cbp.gov/newsroom/local-media-release/rgv-agents-deploy-life-saving-rescue-beacons

37.

U.S. Customs and Border Protection. 2021b. “Big Bend Sector Touts Rescue Beacon Effectiveness.” https://www.cbp.gov/newsroom/local-media-release/big-bend-sector-touts-rescue-beacon-effectiveness

38.

U.S. Customs and Border Protection. 2023. Rescue Beacons and Unidentified Remains, Fiscal Year 2023 Report to Congress.”

39.

U.S. Customs and Border Protection. 2024. FOIA request for data from U.S. Customs and Border Protection. FOIA Number: CBP-FO-2024-053704. Data provided to University of Texas at Austin on March 20, 2024.

40.

U.S. Department of Justice. 1998. INS Launches Unprecedented Borderwide Public Safety Initiative. Washington, DC, United States: Press release.

41.

U.S. Department of the Interior. 2022. FOIA request for data from National Park Service. FOIA Number: NPS-2022-001933. Data provided to University of Texas at Austin on March 18, 2022.

42.

U.S. Department of the Interior. 2024. FOIA request for data from U.S. Fish and Wildlife Service. FOIA Number: DOI-FWS-2024-000220. Data provided to University of Texas at Austin on January 5, 2024.

43.

U.S. Government Accountability Office. 2006. “Illegal Immigration: Border-Crossing Deaths Have Doubled Since 1995: Border Patrol’s Efforts to Prevent Deaths Have Not Been Fully Evaluated.” Report No. GAO-06-770.

44.

U.S. Government Accountability Office. 2022. “Southwest Border: CBP Should Improve Data Collection, Reporting, and Evaluation for the Missing Migrant Program.” Report No. GAO-22-104374.

45.

Williams

J. M.

2013. “When Arrest Becomes Rescue: Border Enforcement and the Politics of Humanitarianism at the US-Mexico Border.” PhD dissertation, Clark University, Worcester, Massachusetts.

46.

Williams

J. M.

2016. “The Safety/Security Nexus and the Humanitarianisation of Border Enforcement.” The Geographical Journal 182 (1): 27–37. https://doi.org/10.1111/geoj.12119.

47.

Winograd

2005. “O’odham Land Gets 8 Rescue Beacons.” Tucson Citizen 1.

48.

Yorkville Enquirer. 1912. Wily Smugglers of Chinese. Yorkville Enquirer.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.93 MB