Essential Air Service in the United States

Abstract

The Essential Air Service (EAS) program continues to receive federal funding to provide air travel access to rural communities in the United States. Regular assessment and evaluation of the performance of this program are important, given limited federal resources and fiduciary duties. In fact, this program has garnered significant attention through the years because it was originally conceived to offer temporary financial bridging for maintaining commercial air service in rural and remote communities following deregulation in 1978, yet has continued to be funded at increasing rates for over thirty-five years. This article undertakes a systematic analysis of the EAS program in terms of access and intended goals and objectives. A spatial optimization model is used to examine service performance of the existing system. Program insights as well as ways system efficiency that could be enhanced are highlighted for rural air transportation service in the United States.

Keywords

rural development planning economic growth and development transportation and transit urban and regional issues spatial structure location models spatial analysis geographic information science optimization cost–benefit analysis

Introduction

One of the more controversial federal transportation programs in the United States is Essential Air Service (EAS; Grubesic, Murray, and Matisziw 2013). The EAS program allocates federal subsidies to commercial air carriers for the purpose of providing air service to rural and remote communities in the United States (US Department of Transportation [USDOT] 2013). Originally conceived as a “safety net” for communities that would be abandoned by commercial carriers seeking to serve larger cities with more profitable routes, the EAS program has continuously operated since 1978 with increasing levels of subsidy. The basic rules for EAS eligibility are twofold. First, the community seeking a subsidy is required to have had commercial air service prior to deregulation in 1978. Second, the community must be located more than seventy highway miles from the nearest Federal Aviation Administration (FAA)-defined medium or large airport hub (USDOT 2009).¹ The geographic criterion has remained a constant since US Congressional reforms to the program in 1990. However, in an era of shrinking budgets and increased accountability, new reforms to the EAS program have been designed to make the system more efficient. For example, the “Airport and Airway Extension Act” (Public Law No. 112-27) prohibits EAS support in communities “where the annual per-passenger subsidies exceed US$1,000, regardless of their distance to the nearest hub airport” (USDOT 2013, 1).² The “Consolidated and Further Continuing Appropriations Act, 2012” (Public Law No. 112-55) also removed requirements that air service be provisioned on fifteen seat or larger aircraft. The recent “Federal Aviation Administration (FAA) Modernization and Reform Act of 2012” (Public Law No. 112-95) permanently reduced the set of communities eligible for EAS support, only allowing those participating in EAS between September 30, 2010, and September 30, 2011, to be eligible (USDOT 2013). In effect, this means that no new communities may be added to the EAS program, even if they lose service, although this is not necessarily binding.³ Finally, beginning in fiscal year 2013, EAS-subsidized airports are generally expected to maintain an average of ten passenger enplanements per service day in the contiguous United States. Exceptions are made for Alaska, Hawaii, and communities that are more than 175 highway miles from the nearest large or medium hub.

While these additional regulations have helped the EAS program to evolve, issues remain regarding spatial and operational inefficiencies of the subsidized system (US Government Accountability Office [GAO] 2009). Research has sought to characterize efficiency and inefficiency of the EAS program in various ways. One approach involves comparing the ratio of federal subsidy (input) and passenger enplanements and/or passenger load factors (PLFs; output). Previous work in this area suggests that some EAS airports operate more efficiently than others (Grubesic and Wei 2012). Another approach has focused on spatial efficiency of the EAS airport system. This has involved assessing costs associated with locational arrangement compared to costs associated potential alternative arrangements of a system. Research suggests that eligibility requirements could be modified to reduce spatial inefficiencies inherent to the EAS program (Grubesic and Matisziw 2011; Grubesic, Matisziw, and Murray 2012; Grubesic, Murray, and Matisziw 2013) and that routes between EAS airports and connecting hubs could be optimized to reduce costs and improve program performance (Matisziw, Lee, and Grubesic 2012).

The purpose of this article is to build upon the growing work pertaining to rural air transport in the United States, with a specific focus on the EAS program. Previous work has focused on spatial efficiency (Grubesic, Matisziw, and Murray 2012; Grubesic, Murray, and Matisziw 2013) and operational efficiency (Grubesic and Wei 2012) of EAS airports but has not examined the interplay of these different types of efficiency. This is a significant gap in the literature and conceptualization of performance efficiency. First, spatial and operational efficiency are often at odds with each other, meaning that some type of trade-off could be beneficial for balancing coverage of market demand and efficiency of carrier operations. Second, efforts to explore trade-offs require specifically tailored modeling approaches. Third, as detailed previously, the EAS program is highly dynamic, with yearly changes to the number of communities included, passengers served and subsidies allocated. As a result, EAS is a moving target. This article benefits from the use of more geographically detailed and updated information, incorporating the most current publicly available data for the EAS program. Thus, although there has been a flurry of recent work on EAS, the data, methods, and results of this article break new ground.

Background

There has been considerable interest in air transport service provision to rural communities in North America, South America, Asia, and elsewhere (Kanafani and Abbas 1987; Kaemmerle 1991; Reynolds-Feighan 1995; Goff 2005; Reynolds-Feighan and McLay 2006; Derudder, Witlox, and Taylor 2007; Fenley, Machado, and Fernandes 2007; Grubesic, Matisziw, and Zook 2008; Grubesic, Matisziw, and Zook 2009; Tierney and Kuby 2008; Ishii, Jun, and Van Dender 2009; Matisziw and Grubesic 2010; Jin et al. 2010). Although air transportation is generally viewed as a technology that compresses both space and time (Zook and Brunn 2006), most large countries, including the United States, have had a difficult time serving their broad geographical expanse with an adequate number of access points to the commercial air transport system. As of October 2013, there were nearly 20,000 airports, heliports, and seaplane bases in the United States (FAA 2013). However, only a few hundred of these air transport facilities actually offer regular commercial service. The commercially served airports in the United States are shown in Figure 1.⁴ There are massive costs associated with operating this commercial air transportation system, and only a limited number of facilities can accommodate larger aircraft and/or undertake comprehensive maintenance. The topological structures of commercial airline networks (e.g., hub-and-spoke) are designed by service providers to promote efficiency, minimize disruptions due to weather, and exploit economies of scale and scope (O’Kelly and Miller 1994; Goetz and Sutton 1997).

Figure 1.

Commercial airports in the United States, 2011.

The EAS program occupies a somewhat awkward niche in the air transport landscape in the United States. As detailed previously, EAS was established to ensure that rural and remote communities maintained access to the national air transport system when (and if) commercial carriers found such communities to not to be economically viable to serve. Today, although commercial carriers have migrated to more profitable routes (Grubesic, Matisziw, and Zook 2008), this process did not leave every small, rural, or remote community in the United States dependent on federal subsidies to maintain commercial air service. For example, the Klamath Falls Airport (LMT), located in Southern Oregon, maintains commercial service via SkyWest Airlines with connections to Portland (PDX) and San Francisco (SFO). With a population of 20,840 (Census 2010), Klamath Falls is not dramatically different from many EAS communities in the United States, but a combination of factors clearly make subsidies unnecessary. The precise mix of demographic, socioeconomic, business, and locational factors that facilitate a successful and profitable service relationship between commercial carriers and the communities they serve remains unclear (Grubesic and Wei 2013). A similar conclusion is true for EAS-subsidized communities. The factors that drive passenger demand at an EAS airport in Wisconsin may be completely different from those factors found in Arizona, Alabama, or elsewhere.

Today, service utilization in the EAS system remains extremely low for many subsidized routes (Grubesic, Murray, and Matisziw 2013). For example, the route between Ely, Nevada (ELY), and Las Vegas (LAS) was subsidized at a rate of US$1637.45 per passenger (2012), far exceeding the US$1,000 per passenger limit set by the USDOT (Eggen 2012). Empty aircraft on this route and many others (see Yamanouchi 2010) has generated severe criticisms of programmatic costs and government waste in an era of sequestration and federal shutdowns (Plumer 2013).

EAS Performance

Recent work by Grubesic, Murray, and Matisziw (2013) details four key measures used by the federal government to evaluate the performance of the EAS program (see also White House 2009). The two outcome-oriented measures track how effectively the government renews carrier contracts prior to expiration and if service is continuously maintained at a level of 98 percent for the entire network. The output-oriented measure tracks the percentage of new contracts processed within 160 days. Finally, the only measure of efficiency tracks the percentage of payments to air carriers that are processed within fifteen business days. As noted by Grubesic, Murray, and Matisziw, there is a notable gap in measuring EAS performance because none of the federal measures account for passengers served, flight frequencies, load factors, itinerary pricing, or many of the other factors that have a dramatic impact on the viability of commercial air service in rural communities.

Operational Efficiency

One important way to measure the efficiency of EAS operations is tracking PLFs on served routes (Schefczyk 1993). A PLF is a measure of how much carrying capacity is used. Basically, the PLF reflects the number of passenger miles flown as a percentage of seat miles available. If the flight distance between two cities is 700 miles and the aircraft has fifty total seats with only twenty occupied by paying passengers, the load factor is 40 percent. Generally speaking, as PLF values approach 100 percent, commercial operators are content because there is little excess capacity in the system. As PLF values approach 0 percent, excess capacity exists. In 2011, there were 116 EAS-subsidized airports in the United States. The average load factor for all carriers across this system was 41.45 percent (Figure 2).⁵ This suggests that a significant level of excess capacity exists in the EAS system.

Figure 2.

EAS airport load factors, 2011.

An issue with the PLF measure, however, is that it lacks context. This is particularly true when considering the EAS program because each carrier (and community) receives a different subsidy. That is, all EAS airports are not treated uniformly. In addition, each community has a different competitive backdrop—some are relatively close to medium or large hubs, while others are farther away. As a result, the fundamental question regarding operational efficiency still pertains to PLF performance, but is shaped by subsidy levels and the local competitive landscape.

Spatial Efficiency

A second important way to measure EAS efficiency is by evaluating the spatial distribution of program operations (Flynn and Ratick 1988; Grubesic and Matisziw 2011). As mentioned previously, spatial efficiency refers to the process of assessing costs associated with a given locational arrangement of a service and comparing them to costs associated with the best-known alternative arrangement (Murray 2001; Desai and Storbeck 1990). Where EAS is concerned, this evaluation can be accomplished through spatial optimization models (Grubesic, Matisziw, and Murray 2012; Grubesic, Murray, and Matisziw 2013). Such approaches enable assessment of the degree of redundant geographic service and marginal benefits for reduced levels of service.

Combined Efficiency

Considering that both operational and spatial efficiency are critical to the health of the EAS network, and the federal program more generally, the challenge for effectively evaluating EAS efficiency is accounting for both operational and spatial components. For the purposes of this article, a loosely coupled suite of optimization-based techniques are used for generating a more holistic snapshot of efficiency. Specifically, we combine PLF scores with several key characteristics for the market areas of each EAS community in a data envelopment analysis (DEA) model (Charnes, Cooper, and Rhodes 1978; Banker, Charnes, and Cooper 1984) for determining technical efficiency. Once the most operationally efficient airports are determined using DEA, the derived efficiency scores are utilized in a spatial optimization model to evaluate trade-offs between maximizing potential demand coverage and maximizing operational efficiency in the EAS network. As noted previously, this facilitates determining if there are EAS network configurations that provide more “bang for the buck,” reducing the financial burden to tax payers and making the EAS program more viable in an era of budget reductions and financial accountability.

Method

Evaluation of EAS system performance and efficiency relies on the use and integration of several databases. Commercial airport data for 2011 were obtained from the National Transportation Atlas (Bureau of Transportation Statistics [BTS] 2011a). These data contain all the locations of public-use aviation facilities for the United States and were combined with data from BTS (2011b) on PLFs. These data were reduced to include only airports with commercial service for 2012. EAS subsidy rates, current as of May 2011, were also obtained from the USDOT (2012). Block group population estimates for 2011 were obtained from Environmental Systems Research Institute (ESRI 2011) to derive the underlying population within the seventy-mile network catchment area around each airport using ArcGIS.⁶ Road network data were acquired from ESRI. Distance assessment was based on shortest network travel.

Operational Efficiency

As noted previously, operational efficiency can be assessed using DEA. DEA is rooted in early work on production functions and efficiency (Farrell 1957). Specifically, DEA is a nonparametric technique that measures the relative efficiency of a peer unit, or decision-making unit (DMU), in the utilization of resources compared to output (Charnes, Cooper, and Rhodes 1978; Banker, Charnes, and Cooper 1984). Identification of an efficiency frontier and associated efficiency scores for DMUs are derived using an optimization model (Charnes, Cooper, and Rhodes 1978; Desai and Storbeck 1990).

In the context of EAS operational efficiency, airports serve as the units of analysis, or the DMU. PLFs function as the output of interest. The input variables include a measure of potential demand (seventy-mile catchment area population), two measures of local competition (distance to medium and large hubs), and the subsidy allocation for each airport. These model inputs are widely used for statistically evaluating demand for commercial air service in rural and remote communities (Kanafani and Abbas 1987; Kaemmerle 1991; Goff 2005; Grubesic and Wei 2012). For typical commercial airports, subsidy levels are not a consideration, but for the EAS program they are a critically important aspect of local service. Using a DEA output-oriented approach (see Banker, Charnes, and Cooper 1984; Grubesic and Wei 2012), the primary output of the analysis is an efficiency score. The DEA data domain consists of a data matrix (Seiford and Zhu 2002):

P = [\begin{matrix} Y \\ - X \end{matrix}] = [P_{1}, . . ., P_{n}],

with s + m rows and n columns, with each column corresponding to one of the EAS airports. Let j represent the index of EAS airports (DMUs), so that $P_{j} = [\begin{matrix} Y_{j} \\ - X_{j} \end{matrix}]$ is composed of an input vector x _j whose ith component x _ij is the amount of input i used by airport j. Efficiency scores, s _j, are obtained through a linear programming problem (Banker, Charnes, and Cooper 1984) for each airport j. The efficiency score for each airport ranges from 0 to 1. Scores of 1.0 indicate airports that are performing the best in relation to their peers based on inputs and outputs. Such airports are operationally efficient in the sense of maximizing load factors, given their subsidy rates, local potential demand, and competition. As a technical efficiency score trends to 0, the airports are less operationally efficient. The types of inefficiencies possible in this framework include insufficient local potential demand, too close in proximity to competing hubs, excessive subsidy rate, or underperformance given local resources.

Spatial Efficiency

The spatial efficiency of the EAS network, or potential changes to that network, is assessed using a spatial optimization model. This is a coverage-based approach related to the maximal covering location problem originally proposed by Church and ReVelle (1974). Differing from previous coverage-oriented analysis of EAS systems by Flynn and Ratick (1988), Grubesic, Matisziw, and Murray (2012), and Grubesic, Murray, and Matisziw (2013), our approach seeks spatial configurations of the EAS system that maximize potential demand coverage (spatial efficiency) as well as total DEA-based scores (operational efficiency), but do so while maintaining a limit on total EAS subsidy allowed. Effectively, we are interested in whether there is any spatial inefficiency in the existing system, and if so, where this is occurring. Consider the following notation:

i = index of block groups to be served;

j = index of EAS airports (DMUs);

Ω _i = set of airports that can serve block group i;

s _j = efficiency score of airport j;

m _j = federal subsidy given to airport j;

Ψ = total subsidy budget;

g_i = population in block group i;

$X_{j} = \{\begin{matrix} 1, i f a i r p o r t j i s s e l e c t e d t o r e m a i n i n t h e E A S s y s t e m \\ 0, o t h e r w i s e \end{matrix}$

$Y_{i} = \{\begin{matrix} 1, i f b l o c k g r o u p i i s c o v e r e d b y a m a i n t a i n e d E A S a i r p o r t \\ 0, o t h e r w i s e \end{matrix}$

With this notation, coverage model specification is possible for the approach utilized to evaluate spatial efficiency.

M a x i m i z e \sum_{i} g_{i} Y_{i} .

M a x i m i z e \sum_{j} s_{j} X_{j} .

S u b j e c t t o \sum_{j \in Ω i} X_{j} \geq Y_{i}, \forall i .

\sum_{j} m_{j} X_{j} \leq Ψ .

X_{j} = {0, 1}, \forall_{j} .

Y_{i} = {0, 1}, \forall_{i} .

Objective (2) seeks to maximize potential demand covered by EAS airports that are to remain in the system. Objective (3) seeks to maximize the operational efficiency of the airports that remain in the system. Constraint (4) tracks whether passenger demand in block group i is suitably served by one or more airports that remain in the system. Constraint (5) represents a system budget constraint, limiting the number of EAS airports that can receive a subsidy. Constraint (6) specifies integer requirements on decision variables.

Where the variables are concerned, it is important to remember that the index set of block groups corresponds to locales that are not already covered by medium or large hub catchment areas. Thus, when given a choice between EAS or medium/large hub service, it is assumed that passengers will be drawn to hubs and that block group is excluded from the index set.

The issue of interest in this article, addressing operational and spatial efficiency, is unique. As a result, accurately representing issues of operational efficiency, spatial efficiency, and budget limitations requires some variation to classic coverage model approaches. The formulation is most closely related to the work of Church and ReVelle (1974). Some important differences include a second function to be optimized, objective (2), as well as the use of a budget threshold, constraint (4), instead of a specification of exactly p facilities to site. Flynn and Ratick (1988) included cost as a second objective (as well as stopover), with a goal of minimizing total costs. Other coverage models have in fact included generalized cost constraints similar to equation (4) (see Moore and ReVelle 1982). An accurate characterization of the previously mentioned model is that it represents a variant of the work of Church and ReVelle (1974).

As noted previously, this model represents an approach that can be used to evaluate the efficiency of a service system that is inherently spatial. In this case, the issue is providing air service coverage to rural communities, so a network of airports that can efficiently carry this out is desirable. Further, this must be done under circumstances that dictate limited subsidy potential and achievement of the greatest system efficiency possible. The proposed model enables these particular issues to be assessed in the context of the existing EAS network of airports, providing insights into possible inefficiency as well as guidance in the case of likely reduced program subsidy in the future.

Data

The EAS network in the continental United States was considered. This involved 116 airports (see Figure 1) that uniquely serve 30,635,878 people in 24,034 block groups, based on 2011 program standards and guidelines.⁷ In particular, the served population is defined according to a seventy-mile (network travel distance) catchment. ArcGIS was utilized for spatial data management, integration, and processing.

EAS airport operational efficiency was derived using DEA. This was done using Frontier analysis, a commercial software package (Banxia 2012). Four inputs (potential demand based on a seventy-mile catchment area, distance to medium hub, distance to large hub, and subsidy allocation) and a single output (PLF) for each airport were considered.

Spatial efficiency was assessed by applying the previously mentioned coverage optimization model for a range of funding scenarios. Implementation of the model was coded in Python. Gurobi, a commercial optimization software package, was then used to obtain optimal solutions for each problem instance. The weighting method (Cohon 1978) was employed to identify nondominated solutions reflecting the trade-off of operational and spatial efficiency. Solution times for all problem instances were less than one second.

Results

For the 2011 budget of US$216,704,578, all 116 airports remain in the EAS network. The total operational efficiency is 8,795.07 using the DEA scores for each airport. This corresponds to an average operational efficiency of 75.81 percent (8,795.07/116). The spatial efficiency is 100 percent because all potential demand for service (30,635,878 people across 24,034 block groups) is within the EAS program-stipulated service distance of seventy miles of an airport. The major question to be addressed is which airports should/could be eliminated from the system in order to have the least possible impacts on the spirit and intent of the EAS program? From an interpretive perspective, improvements in average operational efficiency reflect an enhanced use of resources. Reduced levels of potential demand served represent degradation of the EAS system coverage. Of fundamental interest then is whether there are scenarios that can be identified that reflect a better use of resources while impacting the fewest people possible.

To begin, consider the situation where only 90 percent of the 2011 budget is available for the EAS program, reducing total funding to roughly US$195,034,120. Practically speaking, this means that some of the currently subsidized airports will need to be eliminated from the program as nearly US$22 million in reductions must be absorbed by the system. Figure 3 illustrates the trade-offs derived using the spatial optimization model.⁸ With an emphasis on spatial efficiency, objective (1), it is possible to serve 99.7 percent of all potential demand within the seventy-mile EAS program standard using only 104 airports. This is suggestive of very little impact on the rural population served when twelve airports in the EAS program are eliminated. The targeted airports for elimination from the system are indicated in Figure 4. These changes to the system mean that a reduction in the EAS program operating budget of US$21,772,786 is possible without significantly degrading spatial coverage. Couched somewhat differently, the incremental cost associated with serving only 0.3 percent of potential rural demand, or 92,103 people, equates to US$236.39 per person. Because the actual utilization of the system is likely to be far less than this number of people, the subsidy per passenger would increase substantially. Operational efficiency for this particular scenario is 7,975.68 (objective 2). This represents a slight improvement in average operational efficiency, 76.68 percent (+0.87 from current system).

Figure 3.

Trade-off curve for spatial and operational efficiency at a 90 percent budget level.

Figure 4.

Spatial efficiency gains via a 10 percent budget reduction (w = 1).

If the emphasis is shifted to placing more importance on operational efficiency (objective 2), differences in the spatial distribution of maintained EAS airports can be observed. For this case, 105 airports are to be preserved and would serve/cover 93.12 percent of the potential demand (rural population). This spatial configuration is shown in Figure 5. Total operational efficiency is 8,520.69, which is an average operational efficiency of 81.14 percent (+5.33 from current system). To achieve this, however, the amount of potential demand covered by the system declines. The trade-off that emerges, therefore, is that an increase in operational efficiency is only possible by a reduction in spatial efficiency, objective (1). As a result of this trade-off, multiple airports in Montana and Kansas are now targeted for removal from the EAS system in order to meet budget reductions. The reason these airports are identified is that their average PLFs for these airports is relatively low, as is their average operational efficiency. In Montana, for example, the average PLF is about 11 percent and the average operational efficiency score is 14.33 percent. These are among the lowest in the system.

Figure 5.

Operational efficiency gains via a 10 percent budget reduction (w = 0).

As suggested in Figure 3, there are also seven other trade-off solutions between the previously mentioned two extremes. A more balanced trade-off is where both spatial and operational efficiency are given equal emphasis. The recommended configuration in this case is identified in Figure 6, illustrating the 106 airports that are to remain in the system. This case results in 98.7 percent of the potential demand being served/covered and a total operational efficiency of 8,478.29. This equates to an average operational efficiency of 79.98 percent (+4.17 from current system). Achieving this more balanced trade-off between spatial efficiency and operational efficiency results in a configuration of airports that would substantially impact the northern tier of the United States (most of Montana; Devils Lake, SD; Thief River Falls, MN; Presque Isle, ME). Both Dubois, Pennsylvania, and Liberal, Oklahoma, would also lose service. Once again, the spatial and operational efficiencies for many airports in Montana are relatively poor, so it is no surprise that they are targeted for elimination from the system. Dubois, PA (DUJ) is also recommended for removal. In addition to its load factor (29.08) and operational efficiency (46.44 percent) being relatively low, it is sandwiched between four other EAS airports (FKL, BFD, JST, and AOO), all of which likely compete for potential Dubois passengers on the edge of its seventy-mile catchment area. This process of cannibalization is not uncommon among geographic clusters of EAS airports (Grubesic, Matisziw, and Murray 2012).

Figure 6.

Combined efficiency gains via a 10 percent budget reduction (w = 0.5).

In an era of significant and sustained budgetary cutbacks, a reduction of only 10 percent may not be realistic. Of course, using the spatial optimization model any reduced level of subsidy could be considered. A sampling of results is provided in Figure 7, showing the trade-off curves for six different budgets, ranging from 50 percent to 100 percent of the existing US$216,704,578 allocated to the EAS program in 2011. The 100 percent and 90 percent cases discussed previously are included to provide comparative context. It is interesting to note that for the 80 percent case, all trade-off scenarios are still capable of serving over 90 percent of potential rural demand currently served by the system. In fact, 98.77 percent of potential demand remains served by only ninety-four airports when emphasis is placed on spatial efficiency. This yields a fiscal savings of US$43.5 million for the year. The operational efficiency in this scenario is 7,342.82, an average of 78.11 percent (+2.3 from current system). When operational efficiency is emphasized for the 80 percent case, more airports are required (ninety-seven) and less demand is covered (90.79 percent), but the average operational efficiency increases dramatically to 83.97 percent (+8.16 from current system). The more balanced treatment of operational and spatial efficiency for the 80 percent case results in 96.22 percent of the potential demand being covered by ninety-seven airports with an average operational efficiency of 83.20 percent (8,070.85 total efficiency in Figure 7).

Figure 7.

Trade-off curve for spatial and operational efficiency at multiple budget levels.

Again, similar trade-offs exist for any level of budget reduction. The 40 percent case yields fairly significant level of financial savings (over US$86 million), without severe decreases in spatial efficiency. For example, an interesting trade-off is the scenario where seventy-nine airports are included in the system, capable of covering 82.3 percent of the potential rural demand. The average operational efficiency is 89 percent (total operational efficiency is 7,074.02 in Figure 7). This is indicative of a highly efficient configuration of airports in terms of either operational or spatial efficiency measures.

Discussion

The analyses based on DEA (operational efficiency) and potential demand for service (spatial efficiency) are compelling. Using the spatial optimization model, it is possible to identify a range of trade-off solutions for any stipulated level of subsidy reduction. For the cases examined here, it can be argued that EAS system reconfiguration is possible while keeping spatial impacts to a minimum, and in doing so average operational efficiency can be significantly enhanced.

An interesting issue to consider is what would happen without the use of the spatial optimization approach detailed here in the situation where multiple communities in the EAS program are being targeted for termination. One example of this is an order issued by the US Department of Transportation (2004), where a review of the passenger traffic for seven communities was conducted. Two facets of airport performance were considered: (1) passenger traffic and per-passenger subsidy and (2) distance to nearest medium or large hub. The review (USDOT 2004, 2) suggested that:

Subsidy at three of the communities (Enid, OK, Ponca City, OK, and Brownwood, TX) exceeds the statutory ceiling of $200 per passenger; that those three communities are located within 210 miles of the nearest large or medium hub; and that, consequently, they are no longer eligible for subsidy to support their scheduled service

Simply put, the USDOT used a combination of passenger statistics, subsidies, and distance to nearest hubs to determine potential subsidy reduction through the elimination of airports in the system. This is consistent with the metrics used for evaluation in this article. A shortcoming of the previously mentioned USDOT approach, however, is that it stresses costs reductions due to operational inefficiencies without considering potential systemic impacts (e.g., spatial efficiency) of airports in the system.

Consider for a moment the use of the two core facets of the USDOT strategy applied to the entire EAS system. The implication is that airports would be ranked based on passenger traffic and subsidy. Imagine then a naive approach that merely targets poor operational efficiency airports, essentially ignoring spatial context and spatial efficiency. As detailed by Grubesic and Matisziw (2011) and Grubesic, Matisziw, and Murray (2012), there are many instances where EAS airports are actually cannibalizing each other, are too close to a small hub, or serve a special geographic purpose. The result of a sorting and removal of poorly performing airports based on passenger traffic and subsidy is shown in Figure 8 that would achieve a 10 percent reduction. Suggested for elimination are both Yellowstone Regional Airport (COD) in Cody, Wyoming and Yellowstone Airport (WYS) in West Yellowstone, Montana. The problem here is that both airports are considered technically efficient by the DEA model and that the sort/remove method fails to account for the more nuanced aspects of operational efficiency in the context of the overall system. Because tourism is a primary driver for both of these markets, limited criteria and local catchment area characteristics do not tell the full story for WYS or COD. In comparative terms, this approach can be compared to the trade-off alternative already discussed for the 90 percent funding case (Figure 3). The sort/remove configuration of airports is evaluated with respect to objectives (1) and (2), and then displayed in Figure 9 in relation to the model derived trade-off curve. Three cases of sort/remove are considered: operational efficiency score, population served (spatial efficiency), and a weighted combination of these two. Interestingly, all three configuration identified are inferior. From an economic perspective, all three of these solutions are dominated, as they lie within the Pareto efficiency curve. These solutions serve to highlight that the use of the optimization-based approach for exploring EAS efficiency is essential, providing powerful insight for policy and planning concerning programmatic performance improvements.

Figure 8.

Efficiency gains generated by a sort and remove approach (traffic and subsidy) at 90 percent of total budget.

Figure 9.

Trade-off curve for spatial and operational efficiency at a 90 percent budget level with sort/remove solutions.

Conclusion

EAS is one of the most criticized and heavily monitored federal programs in the United States. As recently as July 2013, the US House of Representatives rejected two separate attempts to either completely eliminate EAS or reduce subsidy levels. For example, in a 166–248 vote, Rep. Tom McClintock’s (Republican, California’s 4th District) proposal to completely terminate the EAS program was rejected. In a separate vote, Rep. Alan Grayson’s (Democrat, Florida’s 9th District) proposal to reduce top per ticket subsidies from US$500 to US$250 was also voted down, even though it had majority support from Republican lawmakers (pp. 137–88).⁹ This level of scrutiny is somewhat normal for the EAS program. Amazingly, even with a constant stream of negative attention, the EAS program continues to flourish, with an operating budget that grew from US$96.5 million in April 2007 to levels exceeding US$220 million in October 2012 and February 2013. This level of growth is even more astonishing when one considers that the United States is currently operating with a federal debt of nearly US$17 trillion and the government is shut down (RT.com 2013).

At the same time, there is no doubt that access to the commercial air transport system remains challenging, yet likely needed in a number of ways for rural and remote areas in the United States and elsewhere (Reynolds-Feighan 1995). By simply providing a higher level of accessibility from a rural community to larger metropolitan areas, one ingredient for economic growth is present (Keeble, Owens, and Thompson 1982; Keane 1984; Goetz 2002). Further, not only does EAS provide a critical transportation link for rural communities to the outside world, the subsidized airfares facilitate an ease of movement that is not often found in midsized communities where ticket prices are higher (Goetz 2002; Grubesic and Zook 2007). These “pockets of pain” (Goetz 2002) emerge for a variety of reasons: geographic isolation, weak markets, carrier monopolies, and predatory behavior. Regardless of the specific problem, these issues generally serve to slow the degree of spatial interaction between places.

The results of this article suggest that a balance can be achieved for the EAS program. In particular, rather than simply saying that the EAS program should be terminated because it is a waste of taxpayer’s money (Gillies 2000; Sparks 2007; Faler 2009; McClintock 2013) or that the budget should be slashed because of its inefficient use of resources (White House 2006; GAO 2009), our results suggest that a combination of spatial and operational efficiency measures can be used to enhance the EAS program and make it a more financially responsible federal investment without sacrificing the program’s goals and objectives. More importantly, the empirical results of this article suggest that a range of potential reconfiguration scenarios are possible, each offering a viable arrangement of subsidized EAS communities and each offering a unique blend of spatial and operational efficiencies. Although there is a slight decline in the population covered for all of the alternative scenarios presented for the case of the 90 percent current budget (Figure 3), the combination of spatial, operational, and financial savings are compelling. In previous work, Grubesic, Matisziw, and Murray (2012) and Grubesic, Murray, and Matisziw (2013) suggest that extending the eligibility criteria to eighty or ninety miles could generate significant savings as well. While this is true, this article details an alternative approach that does not require policy adjustment. Instead, the idea of a more holistic and multifaceted evaluation of current standards and subsidy allocations may be the easiest and most intuitive approach for improving performance and federal accountability.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

References

Banker

R. D.

Charnes

Cooper

W. W.

. 1984. “Some Models for Estimating Technical and Scale Inefficiencies in Data Envelopment Analysis.” Management Science 30:1078–92.

Banxia. 2012. Accessed April 14, 2014. http://www.banxia.com/.

BTS (Bureau of Transportation Statistics). 2011a. “National Transportation Atlas Databases 2013.” Accessed April 14, 2014. http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_atlas_database/2013/index.html.

BTS (Bureau of Transportation Statistics). 2011b. “Passengers: All Carriers—All Airports.” Accessed April 14, 2014. http://www.transtats.bts.gov/Data_Elements.aspx?Data=1.

Census. 2010. United States Census Bureau. Accessed April 14, 2014. http://www.census.gov/2010census/.

Charnes

Cooper

W. W.

Rhodes

. 1978. “Measuring the Efficiency of Decision Making Units.” European Journal of Operational Research 2:429–44.

Church

R. L.

ReVelle

. 1974. “The Maximal Covering Location Problem.” Papers in Regional Science 32:101–18.

Cohon

J. L.

1978. Multiobjective Programming and Planning. Mineola, NY: Dover.

Derudder

Witlox

Taylor

P. J.

. 2007. “US Cities in the World City Network: Comparing Their Positions Using Global Origins and Destinations of Airline Passengers.” Urban Geography 28:74–91.

10.

Desai

Storbeck

J. E.

. 1990. “A Data Envelopment Analysis for Spatial Efficiency.” Computers, Environment and Urban Systems 14:145–56.

11.

Eggen

2012. “Nevada Airport May Lose Essential Air Service, but Optimistic about the Future.” The Ely (NV) Times. Accessed April 14, 2014. http://archives.californiaaviation.org/airport/msg50922.html.

12.

ESRI (Environmental Systems Research Institute). 2011. “Business Analyst.” Accessed April 14, 2014. http://www.esri.com/software/businessanalyst.

13.

FAA (Federal Aviation Administration). 2013. Accessed April 14, 2014. http://www.faa.gov/airports/airport_safety/airportdata_5010/.

14.

Faler

2009. “Small Cities Save Air Links through Guardian Angels in Congress.” Accessed April 14, 2014. http://tinyurl.com/bm3umu8.

15.

Farrell

M. J.

1957. “The Measurement of Productive Efficiency.” Journal of the Royal Statistical Society. Series A (General) 120:253–90.

16.

Fenley

C. A.

Machado

W. V.

Fernandes

. 2007. “Air Transport and Sustainability: Lessons from Amazonas.” Applied Geography 27:63–77.

17.

Flynn

Ratick

. 1988. “A Multiobjective Hierarchical Covering Model for the Essential Air Services Program.” Transportation Science 22:139–47.

18.

GAO (US Government Accountability Office). 2009. “Options and Analytical Tools to Strengthen DOT’s Approach to Supporting Communities’ Access to the System.” Accessed April 14, 2014. http://www.gao.gov/new.items/d09753.pdf.

19.

Gillies

A. T.

2000. “Flying Pork.” Forbes. Accessed April 14, 2014. http://www.forbes.com/2004/01/21/cz_ag_0121beltway.html.

20.

Goetz

A. R.

2002. “Deregulation, Competition, and Antitrust Implications in the US Airline Industry.” Journal of Transport Geography 10:1–19.

21.

Goetz

A. R.

Sutton

C. J.

. 1997. “The Geography of Deregulation in the U.S. Airline Industry.” Annals of the Association of American Geographers 87:238–63.

22.

Goff

B. L.

2005. “Estimating Determinants of Passenger Air Service to Small Markets.” Urban Studies 42:557.

23.

Grubesic

T. H.

Matisziw

T. C.

. 2011. “A Spatial Analysis of Air Transport Access and the Essential Air Service Program in the United States.” Journal of Transport Geography 19:93–105.

24.

Grubesic

T. H.

Matisziw

T. C.

Murray

A. T.

. 2012. “Assessing the Geographic Coverage of the Essential Air Service Program.” Socio-Economic Planning Sciences 46:124–35.

25.

Grubesic

T. H.

Matisziw

T. C.

Zook

M. A.

. 2008. “Global Airline Networks and Nodal Regions.” GeoJournal 71:53–66.

26.

Grubesic

T. H.

Matisziw

T. C.

Zook

M. A.

. 2009. “Spatio-temporal Fluctuations in the Global Airport Hierarchies.” Journal of Transport Geography 17:264–75.

27.

Grubesic

T. H.

Murray

A. T.

Matisziw

T. C.

. 2013. “A Strategic Approach for Improving Rural Air Transport in the United States.” Transport Policy 30:117–24.

28.

Grubesic

T. H.

Wei

. 2012. “Evaluating the Efficiency of the Essential Air Service Program in the United States.” Transportation Research A 46:1562–73, 21.

29.

Grubesic

T. H.

Wei

. 2013. “Essential Air Service: A Local, Geographic Market Perspective.” Journal of Transport Geography 30:17–25.

30.

Grubesic

T. H.

Zook

. 2007. “A Ticket to Ride: Evolving Landscapes of Air Travel Accessibility in the United States.” Journal of Transport Geography 15:417–30.

31.

Ishii

Jun

Van Dender

. 2009. “Air Travel Choices in Multi-airport Markets.” Journal of Urban Economics 65:216–27.

32.

Jin

Wang

. 2010. “China’s Regional Transport Dominance: Density, Proximity, and Accessibility.” Journal of Geographical Sciences 20:295–309.

33.

Kaemmerle

K. C.

1991. “Estimating the Demand for Small Community Air Service.” Transportation Research Part A 59:101–12.

34.

Kanafani

Abbas

M. S.

. 1987. “Local Air Service and Economic Impact of Small Airports.” Journal of Transportation Engineering 113:42–55.

35.

Keane

M. J.

1984. “Accessibility and Urban Growth Rates: Evidence for the Irish Urban System.” Economic and Social Review 15:125–39.

36.

Keeble

Owens

P. L.

Thompson

. 1982. “Regional Accessibility and Economic Potential in the European Community.” Regional Studies 16:419–31.

37.

Matisziw

T. C.

Grubesic

T. H.

. 2010. “Evaluating Locational Accessibility to the US Air Transportation System.” Transportation Research Part A: Policy and Practice 44:710–22.

38.

Matisziw

T. C.

Lee

C.-L.

Grubesic

T. H.

. 2012. “An Analysis of Essential Air Service Structure and Performance.” Journal of Air Transport Management 18:5–11.

39.

McClintock

2013. “Essential Air Service.” Accessed April 14, 2014. http://mcclintock.house.gov/2013/07/essential-air-service-2.shtml.

40.

Moore

G. C.

ReVelle

. 1982. “The Hierarchical Service Location Problem.” Management Science 28:775–80, 22.

41.

Murray

A. T.

2001. “Strategic Analysis of Public Transport Coverage.” Socio-Economic Planning Sciences 35:175–88.

42.

O’Kelly

M. E.

Miller

H. J.

. 1994. “The Hub Network Design Problem: A Review and Synthesis.” Journal of Transport Geography 2:31–40.

43.

Plumer

2013. “Absolutely Everything You Need to Know about How the Government Shutdown Will Work.” The Washington Post. Accessed April 14, 2014. http://tinyurl.com/kjfo44e.

44.

Reynolds-Feighan

A. J.

1995. “European and American Approaches to Air Transport Liberalization: Some Applications for Small Communities.” Transportation Research Part A 29:467–83.

45.

Reynolds-Feighan

McLay

. 2006. “Accessibility and Attractiveness of European Airports: A Simple Small Community Perspective.” Journal of Air Transport Management 12:313–23.

46.

RT.com. 2013. “US Debt Surges $328 Billion in Single Day, Surpassing $17 Trillion for First Time.” Accessed April 14, 2014. http://rt.com/usa/us-debt-record-default-obama-412/.

47.

Schefczyk

1993. “Operational Performance of Airlines: An Extension of Traditional Measurement Paradigms.” Strategic Management Journal 14:301–17.

48.

Seiford

L. M.

Zhu

. 2002. “Modeling Undesirable Factors in Efficiency Evaluation.” European Journal of Operational Research 142:16–20.

49.

Sparks

2007. “Unfree as a Bird.” The American. Accessed April 14, 2014. http://www.american.com/archive/2007/january/unfree-as-a-bird/.

50.

Tierney

Kuby

. 2008. “Airline and Airport Choice by Passengers in Multi-airport Regions: The Effect of Southwest Airlines*.” The Professional Geographer 60:15–32.

51.

USDOT (US Department of Transportation). 2004. “Order Requesting Proposals, Reducing Authorized Service Levels, and Establishing Final Subsidy Rates”. http://dailyairlinefilings.com/ostpdf51/194.pdf.

52.

USDOT (US Department of Transportation). 2009. “What Is Essential Air Service?” Accessed April 14, 2014. http://www.dot.gov/sites/dot.gov/files/docs/easwhat.pdf.

53.

USDOT (US Department of Transportation). 2012. US EAS Subsidized Report. Accessed April 14, 2014. http://www.dot.gov/office-policy/aviation-policy/essential-air-service-reports.

54.

USDOT (US Department of Transportation). 2013. Essential Air Service. Accessed April 14, 2014. http://www.dot.gov/policy/aviation-policy/small-community-rural-air-service/essential-air-service.

55.

White House. 2006. “Detailed Information on the Essential Air Service Assessment.” Accessed April 14, 2014. http://www.whitehouse.gov/omb/expectmore/detail/10004013.2006.html.

56.

White House. 2009. “Detailed Information on the Essential Air Service Assessment.” Accessed April 14, 2014. http://tinyurl.com/775remc23.

57.

Yamanouchi

2010. “Regional Flights Could Lose Subsidies.” The Atlanta Journal-Constitution. Accessed April 14, 2014. http://www.ajc.com/news/regional-flights-could-lose-864542.html.

58.

Zook

M. A.

Brunn

S. D.

. 2006. “From Podes to Antipodes: Positionalities and Global Airline Geographies.” Annals of the Association of American Geographers 96:471–90.