In this issue:
- Stealth attack on airport privatization
- TSA's flawed study on private screening
- Airport expansion vs. inter-city rail
- Limited profiling endorsed
- Airport competition; why not here?
- News Notes
- Quotable Quotes
Last month Los Angeles City Controller Laura Chick gave the City Council a detailed report on functions the city could outsource or lease, to ease its fiscal crunch. Among those listed was a possible lease of Ontario Airport, one of three airports owned and operated by Los Angeles World Airports. Los Angeles thus becomes the 8th municipal government where public officials have suggested airport privatization during the past year—in addition to Austin, Chicago, Jacksonville, Long Beach, Milwaukee, Minneapolis/St. Paul, and New Orleans.
Two factors have triggered this new burst of interest: Chicago’s success in getting airline approval of its proposed $2.5 billion lease of Midway Airport and the increasingly dire fiscal straits cities and counties find themselves in, thanks to the worst recession since World War II. But there’s one big problem with these officials’ hopes: the federal government. To lease an airport that receives federal Airport Improvement Program (AIP) grants, you have to apply for one of four commercial airport slots in the federal Airport Privatization Pilot Program, which is what permits a municipal airport owner to take lease proceeds “off the airport” and use them for general governmental purposes. And that’s the potential roadblock.
The Pilot Program already provides a number of hurdles for a would-be privatizer, the most significant of which is the two-part airline approval requirement. The proposed deal must win the OK of 65% of the airlines operating at the airport and of airlines representing 65% of the annual landed weight. Chicago met that test in its Midway application, now awaiting final FAA approval this spring. But in the pending FAA reauthorization bill, the House Aviation Subcommittee wants to make this hurdle even higher, increasing the airline approval requirement to 75%. Moreover, the bill would also exclude privatized airports from receiving AIP grants—even though passengers using those airports would still pay the ticket tax which is the principal source of funding for AIP. (Still another problem is that there are now more would-be applicants than there are slots.)
These stealth provisions were quietly inserted in last year’s House bill by Aviation Subcommittee chair Jerry Costello (D, IL), with no debate and no testimony either for or against—my guess is that hardly anyone noticed they were there when last year’s bill was approved by the committee and sent to the House floor, where it was approved. No such provision appeared in last year’s Senate bill, which never made it to the floor. This year’s House bill, HR 915, is nearly identical to last year’s bill. And there in Sec. 143 is the same airport privatization deal-killer language.
It seems odd that a senior Democrat from Illinois, where Mayor Daley is setting the pace on airport privatization, would try to kill the Pilot Program by making it (a) harder to do, and (b) less attractive to airport companies. I’m also dismayed that the Republicans on the Aviation Subcommittee did not try to kill these poison pills. It’s not as if no alternative had been presented. The FAA’s own draft reauthorization bill (in 2007) proposed not only to eliminate the 65% approval requirements but also to expand the number of slots in the Pilot Program to 15.
It’s also odd that House Democrats would want to kill a program that could offer real help to hard-pressed medium and large municipalities, the majority of which are run by Democratic mayors. If these obstructions remain in the House bill (which seems to be on a fast track), the mayors’ only hope will lie in the Senate, where last year’s bill had no provision dealing with airport privatization. Perhaps a city-friendly member of the Senate Commerce Committee’s Aviation Subcommittee will revive the FAA’s proposed liberalization of the program.
In 2007, the Transportation Security Administration awarded a $442,000 contract to Catapult Consultants to analyze the cost and performance of screening at airports with privately contracted screeners versus airports with TSA screeners. Two months after receiving the report in December of that year, the agency issued its own report on the same topic, with its own findings, developed using a different methodology. Neither report has been released, but thanks to an assessment of both by the Government Accountability Office (GAO), we can compare the two studies and speculate about what is going on (see GAO-09-27R).
For its initial cost comparison, the contractor selected six of the 10 airports using private screeners as part of the Screening Partnership Program (SPP) and matched them with six airports of comparable size and functions. Thus, San Francisco (SPP) was compared with Boston Logan (TSA), Rochester (SPP) with Salt Lake City (TSA), etc. This methodology, with an admittedly limited sample size, found that screening costs at the SPP airports averaged 17% higher than at TSA-screened airports. To obtain a larger basis for comparison, the contractor used data from all 450 airports with screening to create a regression model, which it then used to estimate the costs at the six SPP airports if they had TSA-provided screening. By that method, the SPP airports were only 9% more costly. But the contractor also pointed out that TSA maintains potentially redundant administrative staff and overhead at SPP airports, and those costs artificially inflate the estimated cos ts of SPP airports.
In its own report, the TSA used invoices from the private screening companies and internal budget data to conclude that SPP airport screening costs 17.4% more than TSA screening. In its review, the GAO pointed out that TSA left out of its cost comparison such things as workers compensation, general liability insurance, and some retirement costs for TSA screeners paid by the federal government (but not in TSA’s budget), as well as not accounting for the corporate income tax revenue paid to the government by screening companies. TSA also used just one year’s data. GAO also agreed with Catapult Consultants that there are TSA administrative and overhead costs at SPP airports that further distort the cost comparison.
The other dimension of both studies was screening performance. Catapult analyzed four performance measures: threat image projection detection rates by screeners, recertification pass rate of screeners, passenger waiting times, and passenger satisfaction. It used four years of data, again comparing the six SPP airports with the six comparable ones. It found that “SPP airports’ overall performance results are equal to or better than those delivered by non-SPP airports.” TSA’s own study selected five performance measures, but again used data for just a single year, and gave no analysis of the level of confidence in the observed differences between SPP and non-SPP airports. Moreover, GAO does not tell us what TSA concluded about these performance differences, only that it thought TSA’s comparison methodology was weak.
When it comes to comparing costs with performance, GAO notes that TSA did not do this, so it has no basis for deciding whether the reportedly higher costs of SPP screening might be justified by higher levels of performance. And because of the limitations in TSA’s methodology, GAO concluded that “We believe that TSA should not use [their] study as sole support for major policy decisions regarding the SPP.”
So what can we make of all this? It appears to me that TSA received a study from its contractor that revealed inconvenient truths—that SPP screeners performed better and probably at no higher cost than TSA screeners. Rather than release that study to Congress and the media, it quickly did its own study covering the same ground, ignoring some of its own costs and not making clear the results of its own limited performance comparison. And despite the original rationale for having Catapult do the study—which was “to assist senior TSA leadership with strategic decisions regarding the degree to which TSA should leverage public-private partnerships in the area of screening services”—it now tells GAO “that the agency maintains a neutral position on the SPP and neither encourages or discourages airports from applying to participate.”
That ignores what look to me like sensible recommendations from Catapult, namely, that TSA should:
- Reduce its own general and administrative costs at SPP airports;
- Expand SPP to improve performance at low-performance TSA-screened airports;
- Use SPP at “hard-to-hire” airports and airports with large seasonal requirements;
- Explore giving SPP contractors additional “degrees of freedom” to foster innovation, superior performance, and cost controls.
All of those sensible things are within TSA’s purview, under existing law. Let’s hope the new TSA Administrator takes a fresh look at them.
The U.K.’s Labor government has officially approved BAA’s long-standing request to add a third runway at congested Heathrow—but the opposition Conservative party has called instead for a $30 billion high-speed rail (HSR) line between London, Birmingham, Manchester, and Leeds. In Australia, a government study that called for building a second airport for Sydney (where expanding the existing airport is considered politically and socially unfeasible) has been challenged by a proposal from Canberra Airport (149 miles away) to become the second Sydney airport, via construction of a high-speed rail line between the two cities at a cost of about $15 billion. And one of the principal rationales under which California voters approved a $10 billion bond issue last November toward construction of a $50 billion north-south high-speed rail system was to avoid the need to expand airport capacity in both greater Los Angeles and the San Francisco Bay Area.
I doubt that any of these plans makes sense. Consider first of all the distinction between user-pays funding and general taxpayer funding. Airport expansion in the U.K., Australia, and the United States is paid for by airport users, primarily via airport landing fees and space rentals paid by airlines (and recovered from passengers via air fares). By contrast, inter-city high-speed rail’s capital costs are everywhere paid for out of general tax revenues, with operating costs mostly covered out of passenger fares. So the primary public policy question in deciding between airport expansion and HSR should not be “cost” but rather taxpayer cost. If the goal can be accomplished in one case (airport) at zero taxpayer cost and in the other (HSR) only at huge taxpayer cost, then the burden of proof should be on HSR advocates to justify the expenditure of billions of taxpayer dollars.
These days, the principal justification offered is that HSR is “greener” than commercial aviation. While that is often simply assumed by pundits and environmental groups, the data are far from clear. First, the amount of greenhouse gas (GHG) per passenger mile produced by electric-powered HSR depends considerably on the source of electricity. In France, such GHG emissions are low, since most French electricity comes from nuclear power. In the United States and Australia, by contrast, the majority comes from coal.
Second, you need to look at energy consumption per passenger mile. Back in 2004, Prof. Roger Kemp of the U.K.’s Lancaster University compared high-speed rail and airline energy consumption per passenger mile, and found very similar levels for an Airbus A-321 and both the French TGV and the German Inter City Express. More recently, Mikhail Chester and Arpad Horvath of UC Berkeley did a lifecycle analysis of GHG impacts of various transport modes, and found air travel to be environmentally competitive with rail travel.
Third, you need to estimate carefully how much of total airport flight activity might be replaced by a given HSR route. Relatively short-haul trips such as those between London and Birmingham or Los Angeles to San Jose are a relatively small fraction of total activity at major airports such as LAX or Heathrow. In Reason Foundation’s recent “Due Diligence” report on the proposed California HSR system, the authors found that HSR would not significantly reduce the need for additional airport capacity in that state (www.reason.org/ps370.pdf).
Finally, in the case of Heathrow in particular, not expanding its capacity would exacerbate the huge projected airport capacity shortfall in Europe. A Eurocontrol study released in December projected a doubling of passenger traffic by 2030 and the likely saturation of 20 large EU airports. Privatized Fraport recently received final approval to add a fourth runway at Frankfurt, after it reached a deal to pay $858 million to buy up and move a chemical plant located on land needed for the runway. Needless to say, BAA should only build the third Heathrow runway if it pays market value for the 700 homes in Sipson where the new runway would go. And that cost will properly be factored into what airlines and passengers will pay to use the expanded Heathrow.
A key element of risk-based security is the attempt to devote relatively fewer resources to people or locations with low probability of being threats and relatively more resources to those more likely to be threats. This is the premise behind the original concept of Registered Traveler, which some have termed “positive profiling,” since it seeks to create a lower-risk group of passengers by requiring them to pass a background check. But what about identifying the higher-risk group?
That is what secondary screening is all about, to the extent that it is not done solely on a random subset of all passengers. Political debate tends to equate “profiling” with ethnic or religious stereotyping, but even profiling based on statistical risk factors (e.g., paying cash for a one-way ticket) may be open to question on the grounds that the number of potential terrorists is so small compared to the number that match such a profile as to waste a lot of passenger and security system time and resources.
Can mathematics help to resolve this dilemma? The Feb. 10, 2008 issue of Proceedings of the National Academy of Sciences contains a paper suggesting that it can (www.pnas.org/cgi/doi/10.1073/pnas.0813202106). The author is William H. Press, a computational biologist and computer scientist at UT Austin. The paper’s math is beyond me, my calculus having gone mostly unused since my MIT days. So I’m relying here on the New York Times article by Sandra Blakeslee, Feb. 2, 2008, for a simplified explanation of Press’s paper.
The aim is to find an optimal strategy for using prior information to better target screening. The starting assumption is that intelligence agencies can come up with probabilities that individuals meeting a certain profile are more likely to be bad guys. Suppose that certain people are 100 times more likely to be a bad guy than the average passenger. Three alternative policies might be:
1. Always select people with that profile for secondary screening;
2. Select such people 100 times as often as others;
3. Screen them at random, the same as everyone else.
Press’s math concludes that number 1 is the worst option, because it devotes far too many resources to those meeting the profile, with no guarantee that the real bad guy actually matches the profile. Number 2 still subjects far too many innocent people to secondary screening, using scarce security resources. But number 3 misses the opportunity to use the profile data to try to do better than random checking.
Press recommends “square root biased sampling” as the best compromise, somewhere between #2 and #3. In this well-known sampling technique, people matching the profile would be selected for secondary screening 10 times (the square root of 100) as often as people not matching the profile. This method of sampling is often used to identify significant events that might be otherwise lost in a sea of data; Press has used it to find bits of DNA in his computational biology research. He told Blakeslee that it occurred to him when he was passing through airport security that the method could be applied to the profiling problem.
This approach could be incorporated into TSA’s new Secure Flight system, being rolled out this year to replace its antiquated CAPPS (Computer-Assisted Passenger Pre-Screening) system. And it would nicely complement a revised Registered Traveler program that actually did a criminal history background check on applicants, and permitted those who passed to avoid the checkpoint strip and laptop removal so frustrating to frequent flyers
In December, the U.K. Competition Commission recommended that the government break up BAA’s near-monopoly of London and Scottish airports, by ordering it to sell Gatwick and Stansted in London and Edinburgh in Scotland. When the Thatcher government privatized the former British Airports Authority as a single company in 1987, it missed an ideal opportunity to create competition, by selling the airports separately. Competition, rather than price-cap regulation, would have then dealt with the potential of monopoly-like behavior. Assuming the government accepts the recommendation, this historic mistake will belatedly be corrected.
But airport monopoly is not confined to Britain. In the United States, the few metro areas that have more than one major commercial airport generally have only a single airport provider—Chicago, Houston, Los Angeles, and New York being cases in point. The most notable exception is the San Francisco Bay Area, where San Francisco, San Jose, and Oakland are all transcontinental airports but operated by separate, competing entities. Dallas is a partial exception, with Love Field providing a limited degree of competition to DFW. And LAX and Ontario (both operated by Los Angeles World Airports) do have some long-haul competition from Burbank, Long Beach and Orange County.
Is it completely fanciful to imagine antitrust laws ever being applied to U.S. airport quasi-monopolies? This question came to me as I read the powerful new paper by Michael E. Levine, “Airport Congestion: When Theory Meets Reality,” in the current issue of the Yale Journal of Regulation (Vol. 26, No. 1, 2009). The paper is a greatly expanded and far more sophisticated discussion of the subject Levine addressed in his 2007 Reason Foundation policy brief: the real-world difficulties of pricing airport access to address chronic congestion. He expands and elaborates upon his criteria for a new congestion pricing solution, one of which is that airport monopolies must be addressed. (The entire Levine paper is well worth reading, but my focus here is on this single point.)
The problem, exemplified by the three major New York airports, is that “public airport authorities are famously a source of political patronage, vehicles for statements of public grandeur, recipients of Airport Improvement Program pork, and, where exempted, extractors of monopoly rents from airlines.” Thus, Levine fears, correctly, that if such congested airports could charge whatever the market would bear for landing at peak times, they would not be likely to use the proceeds for (politically unpopular) capacity expansion, which is where the proceeds would do the most good (and be in the interest of those paying the higher charges).
While noting that U.K.-type divestiture “is almost inconceivable politically” in the United States (given the political clout of senators and representatives from the urban states where congested monopoly airports are located), Levine follows up with this provocative thought: “Perhaps the financial pressures on municipalities that have made airport privatization more attractive could allow an expansion of the privatization program to be a politically acceptable vehicle for breaking up monopolies, if privatization were made subject to antitrust provisions.” And he notes that “As a result of Midway’s privatization, Chicago airports are no longer a monopoly,” though Midway cannot compete for most international service because its runways are too short.
This is certainly thought-provoking. Would Houston Airport System (which already has a successful privatization subsidiary) consider privatizing Hobby Airport to raise capital to expand that business? Might Los Angeles accept its City Controller’s suggestion to lease Ontario Airport? And might the Port Authority of New York & New Jersey someday contemplate leasing, say, LaGuardia, to raise capital for turning Stewart into the major airport it could become?
Secondary Cockpit Barriers. Strengthened cockpit doors are widely acknowledged as the most cost-effective post-9/11 security reform. But are they sufficient? A small number of pilots has been pushing for the addition of secondary barriers, since (especially on long flights) the cockpit door must be opened several times in flight, for rest-room and food access for the cockpit crew. The FAA is now considering a Part 121 change that would define criteria for secondary barriers. A meeting on this subject takes place in Washington March 3-4, 2009.
BAA Adds Fourth U.S. “Airmall”. BAA USA, which introduced competing brand-name retailers and street pricing to U.S. airport concessions with its original Airmall at Pittsburgh International Airport, has landed its fourth Airmall contract. The new one is a 10-year agreement at Cleveland Hopkins International Airport. Cleveland joins previous BAA operations at BWI, Boston Logan, and Pittsburgh.
NetJets Purchases German Airport. Limited access to slots at congested Frankfurt International Airport has led fractional provider NetJets to purchase nearby Egelsbach Airport, a general aviation airport within the Frankfurt metro area. As a non-scheduled operator, NetJets typically had to submit slot requests just 24 hours in advance of need, and two-thirds of those were turned down, which constrained the company’s growth. Neighboring communities agreed to sell their stakes, since the airport had been losing money, and NetJets agreed to a curfew between 9 PM and 7 AM. It will also invest in upgrading the airport to serve larger business jets.
RAND Terrorism Database. In early February, RAND Corporation announced the availability of the RAND Worldwide Terrorism Incident Database, which encodes over 36,000 incidents recorded over several decades. Access is by subscription only, and 30% discounts are available until April 1, 2009. Details at: www.rand.org/ise/projects/terrorismdatabase
Airport Company Share Prices. The Centre for Asia Pacific Aviation keeps tabs on the share prices of publicly traded airport companies. For example, their December 2008 Airport Investor Monthly included a bar chart showing recent price changes for 21 such companies. Some are airport funds, such as Macquarie Airports, Infratil, and Australia Infrastructure Fund, others are multi-airport operating companies such as Fraport, ASUR, and Airports of Thailand, and some are essentially single-airport companies (Copenhagen, Vienna). In case you’re interested, in that December chart, only six of the 21 showed increases in share price.
“No longer forced to exclusively serve the interests of home-based national carriers, airports now play a key role in driving airline competition. As a result, airport owners and managers across the world are building on this entrepreneurial focus with diversified ownership models including corporatization, private-public partnerships, and full privatization. Such models free airport operators from political interference and red tape, allowing them to focus on the needs of the traveling public and their competitive position.”
--Angela Gittens and Olivier Jankovec, Airports Council International, Letters section, Wall Street Journal, Dec. 9, 2008.
“There has been little speculation about the incoming Administration’s likely attitude towards airport financing, which is yet to register on the radar. . . . U.S. airports, run typically by local councils with private investment by airlines, have become generally unattractive and uneconomic pieces of infrastructure. Airport privatization outside the U.S. has not been universally praised as a concept in action. But certain features have tended to emerge. Where adequate regulation is in place, the quality of infrastructure delivered has generally exceeded the government-delivered standards—especially when the level of government is local.”
--“Airport Privatization,” Airport Investor Monthly, Issue 50, December 2008 (Centre for Asia Pacific Aviation)