Cost-benefit analysis of airport security: Are airports too safe?
Introduction
Much research on aviation security focuses on airplanes due no doubt to the events of September 11 2001 and to the more recent attempts to bomb U.S. bound flights in 2001, 2006 and 2009. Although there may be special reasons to protect airplanes, however, it is not at all clear that there are any special reasons to protect airports. Elias (2010) states that these areas have ‘unique vulnerabilities because it is unsecured’. However, compared with many other places of congregation, people are more dispersed in airports, and therefore a terrorist attack is likely to kill far fewer than if, for example, a crowded stadium is targeted. The 2011 suicide bombing in the baggage claim area of Moscow's Domodedovo airport did kill 37 and injure many others, and this shows that airports are not unattractive targets. However, in the previous year suicide bombers targeted the Moscow metro killing 25, and the year before that, derailed the Moscow to St. Petersburg high-speed train killing 27.
In the fourteen year period 1998–2011, the Global Terrorism Database recorded 20 attacks on airports in the U.S. and Europe, killing 64 people. Notable among these are the attempted bombing of the Glasgow international airport in 2007 and the shooting of two people at the El Al ticket counter at Los Angeles International Airport (LAX) in 2002. Over the same period there were 31 attacks on aircraft. In total, attacks on aviation accounts for only 0.5% of all terrorist attacks, and attacks on airports comprise less than half of these. This experience has led the 2007 U.S. National Strategy for Aviation Security to conclude that ‘reported threats to aviation infrastructure, including airports and air navigation facilities are relatively few.’ A study of 53 cases that have come to light since 9/11 in which Muslim terrorists planned, or in many cases vaguely imagined, doing damage in the United States finds only two in which an airport facility was on the target list (Mueller, 2013).
A risk and cost-benefit assessment quantifies risk reduction of security measures, losses from a successful attack, threat likelihood, probability that attack is successful, and cost of security measures. This allows costs and benefits of security measures to be compared and optimal security measures to be selected. In earlier work evaluating in-flight airline security measures we have considered cost per life saved as the sole decision-support criterion (Stewart and Mueller, 2008), and we later conducted a systems reliability analysis with a more detailed cost-benefit assessment that included other losses from a terrorist attack (Stewart and Mueller, 2013a, Stewart and Mueller, 2013b; see also Jackson et al., 2012). These analyses considered single point estimates of risk reduction and losses. In this paper, we characterise probability of attack success, risk reduction, and losses as probabilistic variables allowing confidence intervals to be calculated (for preliminary efforts, see Stewart and Mueller, 2011). For a literature review of probabilistic terrorism risk assessment see Stewart and Mueller (2013a).
The U.S. Transportation Security Administration (TSA) has extensive security guidelines for airport planning, design and construction (TSA, 2011). However, there is little information about whether TSA guidelines satisfy a cost-benefit assessment. The U.S. Government Accountability Office and Congress have repeatedly urged the TSA to undertake risk and cost-benefit assessments of major programmes (GAO, 2011, Rogers, 2012). The TSA has used the Risk Management Analysis Tool (RMAT) to conduct risk assessments. However, a review by RAND (Morral et al., 2012) revealed a number of key deficiencies. Among them: ‘RMAT does not attempt to describe the absolute risks to the system, rather just the relative risks, or changes in magnitude of risk’, and thus RMAT can only ‘partially meet’ TSA needs. What is needed is a methodology that can assess absolute risk and risk reduction. A key component of assessing absolute risk is including the probability of an attack in the calculations, whereas a relative risk assessment is often conducted conditional on an attack occurring and then ranking risks based on the relative likelihood of threats.
This paper seeks to assess the absolute risks and cost-effectiveness of measures designed to protect airport terminals and associated facilities such as car parks from terrorist attack. These are areas where the general public has unrestricted access to before passengers undertake security screening and pass into secured (sterile) areas prior to aircraft boarding. We rely extensively on cost and risk reduction data for LAX compiled by RAND in 2004 (Stevens et al., 2004), which considered bombings or shooting attacks at the airport curbside or in other pre-screening areas of passenger terminal buildings. We evaluate security measures such as reducing congestion by additional check-in staff and TSA screening lines, making buildings blast-resistant, and screening of vehicles and luggage for IEDs (Improvised Explosive Devices). These range in cost from $2.5 to $60 million per airport per year. LAX is the sixth busiest airport in the world, and third busiest in the United States. Hence, LAX represents a typical large international airport in a class with London Heathrow, New York JFK, and Washington Dulles airports.
The paper first explains risk-based decision theory, and then describes the threats that airport terminal buildings are exposed to, enhanced security measures to deal with these threats, and their cost. The risk reduction for enhanced security measures, loss likelihood, and losses sustained in a successful attack are then inferred. Fatality risks, net present value and benefit-to-cost ratio are calculated for various attack probabilities. The probability of cost-effectiveness is also calculated. This allows the cost-effectiveness of security measures to be assessed and compared, and optimal security measures selected.
Section snippets
Definition of risk
A standard definition of risk or expected loss is:
This is consistent with the conceptual framework adopted by the TSA (NRC, 2010) and risk analyses for many applications (e.g., Kaplan and Garrick, 1981, Stewart and Melchers, 1997). This leads to a simplified formulation for risk:where Pr(T) is the annual threat probability per target, Pr(L|T) is the conditional probability of loss (that the explosive will be successfully detonated or the
Threats, enhanced security measures, and their cost
We consider four significant threat scenarios aimed at airport terminal buildings and associated landside facilities:
T1 large truck bomb – detonated in front of a crowded terminal.
T2 curbside car bomb – detonated in front of a crowded terminal.
T3 luggage or vest bomb – detonated in curbside or inside a crowded terminal.
T4 public grounds shooting attack – terrorists attempt to shoot as many people as possible.
These threats have been called ‘major vulnerabilities’ or ‘major’ threats that can kill
Fatality risks
According to the Global Terrorism Database, in the period 1998–2011 attacks on airport terminals in Europe inflicted 37 fatalities, and 24 fatalities resulted from attacks to airport terminals in the Asia-Pacific area (note that these statistics cover all airports, not just ‘large’ ones). In the same period there was one attack at a U.S. airport (LAX) where a gunmen killed two people at the El Al ticket counter in 2002.
The annual fatality risk is approximately 4.6 × 10−9 for the Asia-Pacific
Conclusions
The risk and cost-benefit decision framework described herein illustrates the key concepts and data requirements. This provides a starting point for this type of risk analysis – and to flesh out some of the issues, including data requirements becoming more challenging as the systems model increases in detail and complexity. Our analysis considered each security measure in isolation, whereas policy options might prefer a mix of security measures. In this case, security measures may also not be
Acknowledgements
Mark Stewart appreciates the financial support of the Australian Research Council.
References (41)
- et al.
Airport security screening and changing passenger satisfaction: an exploratory assessment
J. Air Transp. Manag.
(2006) - et al.
An assessment of the impacts of inspection times on the airline Industry's market share after September 11th
J. Air Transp. Manag.
(2012) - et al.
Cost-Benefit Analysis: Concepts and Practice
(2011) Russian airlines passenger traffic up 12.6%, load factor down in 2011
Air Trans. World
(2012)- et al.
Protecting Commercial Aviation Against the Shoulder-Fired Missile Threat
(2005) Airport and Aviation Security: U.S. Policy and Strategy in the Age of Global Terrorism
(2010)Mitigating risk from abnormal loads and progressive collapse
J. Perform. Constr. Facil.
(2006)Department of Homeland Security: Progress Made and Work Remaining in Implementing Homeland Security Missions 10 Years after 9/11
(2011)- et al.
The economic impacts of a terrorist attack on the U.S. Commercial Aviation System
Risk Anal.
(2007) - et al.
A systems model for probabilistic risk assessment of improvised explosive device attack
Int. J. Intell. Def. Support Syst.
(2012)
Evaluating the societal response to antiterrorism measures
J. Homel. Secur. Emerg. Manag.
Facts and Figures
The Impact of September 11 2001 on Aviation
2012 Annual Review
Efficient Aviation Security
On the quantitative definition of risk
Risk Anal.
Reducing Terrorism Risk at Shopping Centers an Analysis of Potential Security Options
Modeling Terrorism Risk to the Air Transportation System
Quantifying probabilistic expressions
Stat. Sci.
Terror, Security, and Money: Balancing the Risks, Benefits, and Costs of Homeland Security
Cited by (45)
A budget allocation model for domestic airport network protection
2022, Socio-Economic Planning SciencesCitation Excerpt :In literature, a variety of other models are available to enhance the performance of screening systems, for example, cost-benefit analysis models [13–18], simulation models [19–24], decision models [25–29], game-theoretic models [30–33], and fuzzy inference models [34,35]. Besides security screening, researchers have explored other areas in aviation security, for example, the empirical investigation of positive passenger bag matching [36], the threat of man-portable air defense systems [37], checkpoint and canine patrol scheduling [38], federal air marshal scheduling [39,40], and the cost-benefit analysis of different security measures [41–43]. For detailed reviews on operations research applications and risk-based models in aviation security, please refer to Lee et al. [44] and Albert et al. [45], respectively.
Building an assessment model for measuring airport resilience
2021, Journal of Air Transport ManagementMaking sense of airport security in small and medium-sized airports
2021, The Air Transportation Industry: Economic Conflict and CompetitionMulti-criteria group decision-making approach to the modernization of hold baggage security screening system at an airport
2020, Journal of Air Transport ManagementA Stackelberg security Markov game based on partial information for strategic decision making against unexpected attacks
2019, Engineering Applications of Artificial IntelligenceCitation Excerpt :The algorithm for the partially observable random walk is given by: An important real-world application in Stackelberg Security Games is that of protecting airports terminals (Stewart and Mueller, 2014). In spite of compared with many other places of congregation, people are more dispersed in airports, therefore it is an attractive target.
Analyzing airport security checkpoint performance using cognitive agent models
2019, Journal of Air Transport Management
- 1
Tel.: +1 614 2476007.