National Academies Press: OpenBook

Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies (2019)

Chapter: Chapter 2 Survey of Past Analyses of Air Passenger Demand

« Previous: Chapter 1 Introduction
Page 8
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 8
Page 9
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 9
Page 10
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 10
Page 11
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 11
Page 12
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 12
Page 13
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 13
Page 14
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 14
Page 15
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 15
Page 16
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 16
Page 17
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 17
Page 18
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 18
Page 19
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 19
Page 20
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 20
Page 21
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 21
Page 22
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 22
Page 23
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 23
Page 24
Suggested Citation:"Chapter 2 Survey of Past Analyses of Air Passenger Demand." National Academies of Sciences, Engineering, and Medicine. 2019. Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies. Washington, DC: The National Academies Press. doi: 10.17226/25411.
×
Page 24

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

8 C H A P T E R 2 This chapter reviews existing studies of air passenger demand. These studies were undertaken by a variety of individuals and organizations and for a variety of purposes: the relatively pure research motivations of academic researchers, educational or informational purposes exhibited by governmental researchers and reports, or the highly practical objectives of airport profes- sionals and consultants preparing analyses and forecasts of air passenger demand for individual or systems of airports. In this last case, air passenger demand models and forecasts are principal inputs for airport planners and decision makers who must make decisions regarding long lived airport capital investments based in part on uncertain passenger demand forecasts. The purpose of the present research project was to identify and assess options for introducing disaggregated socioeconomic data into efforts to understand and anticipate passenger demand at airports. We used examples of existing practice to identify opportunities for introducing or improving the use of disaggregated socioeconomic data. In the remainder of this chapter, we first summarize past analyses of air passenger demand from academic and government researchers and organizations. We then turn to examples of air passenger demand studies (considered broadly) from past ACRP-managed research. In many of these examples, the authors of the ACRP reports present state-of-the-art standards for practices related to modeling air passenger demand to the ACRP’s readership of airport practitioners and consultants. Surveys of passenger choice are an important source of data regarding how passenger travel choices are related to sociodemographic cohorts and characteristics, and the next subsection provides a survey of the project’s research findings in this area. The subsection is a review of examples from the types of airport documents and studies that are likely to include a treatment of air passenger demand modeling and forecasting. Examples of these documents are airport master plans, airport studies of local passenger demand, and other airport documents in support of airport planning or financing. A final subsection concludes. Academic, Government, and Industry Studies Understanding demand for air travel is an active area of research that has generated a large number of academic, government, and industry studies over the past 50 years that are summa- rized in this section and reviewed in detail in Appendix A. Academic Studies Academic studies of air travel demand vary across multiple dimensions that include: (1) demand measures used as the dependent variable (e.g., modeling the number of passengers on an O&D Survey of Past Analyses of Air Passenger Demand

Survey of Past Analyses of Air Passenger Demand 9 pair or revenue passenger miles); (2) the explanatory variables used as the independent variables, including disaggregate and aggregate socioeconomic variables; (3) the type of model in terms of the functional form used to relate the dependent and independent variables; (4) the type of data (e.g., panel, time series) used for estimation; (5) whether these data were aggregate or dis- aggregated; (6) the market segments used in estimation (e.g., business or leisure travelers); and (7) the time periods on which the models were estimated. Dependent Variables The dependent variable in most air travel demand studies is typically a direct measure of the number of air passengers. The level of route aggregation used to predict the number of air pas- sengers differs across studies. Studies of air travel to or from specific countries, regions, cities, or airports typically consider air passengers without regard to their trip origins (for inbound travel) or destination (for outbound travel). Studies estimating city-pair or airport-pair air travel demand are more common in the academic literature and include Verleger (1972), Ippolito (1981), Fridström and Thune-Larsen (1989), Suzuki and Audino (2003), Bhadra (2003), and Chi, Koo, and Lim (2010). Although the dependent variable in air travel demand studies is typically the number of air passengers, other measures of demand also have been used, including revenue passenger-miles (or revenue passenger kilometers) and travel expenditures. Explanatory Variables A wide range of explanatory variables have been used to model air travel demand. These include variables reflecting different demographic characteristics of travelers, socioeconomic variables that reflect the underlying economic drivers of air travel, measures of air travel service (most notably price), and service measures for competing modes. A variety of aggregate and dis- aggregate demographic variables have been used in the literature to model air travel demand as a function of traveler characteristics. Among recent models, Dargary (2010) included gender, years of residence at the current address, household composition, and type of home in a household-level model of domestic person-miles of travel. Kressner and Garrow (2012) used the proportion of households in selected lifestyle clusters in each zip code to model the average number of home- based air trips per household for zip codes in Atlanta. Population is commonly included in models of air travel demand in city-pair markets (e.g., Castelli, Pesenti, and Ukovich 2003; Bhadra and Kee 2008), although whether the popu- lations of the cities at either end of the market are combined or included as separate variables differs across studies. Population density has also been used in the literature, but tends to be more difficult to interpret. In addition to population, measures of economic activity, such as income or gross domestic product (GDP) are major determinants of air travel demand. Most air travel demand studies (although not all) include some socioeconomic variables that reflect the underlying economic drivers of air travel, although there is little consistency in which variables are used. Income can be measured in a number of different ways (e.g., household income, personal income, total income, disposable income) and at different levels (e.g., for a country, region, or city, or on a per-capita or per-household basis). Of the 45 models reviewed for the project and discussed in Appendix A, 34 used some form of an income measure, including GDP or gross regional product (GRP), per- sonal income or personal disposable income, and household income. In addition to sociodemographic and socioeconomic variables, several air travel demand models have included measures of air service, including average daily flights and measures of air- line market concentration in each city-pair market (e.g., Abrahams 1983; Bhadra 2003). Airfare

10 Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies is the most common measure of air service included in air travel demand studies, although how airfares were defined varied according to the scope of the study. The inclusion of airfare in air travel demand models allows for the calculation of price elasticities. Income elasticities estimated in the different studies varied across studies and across markets and/or market segments within the same study, as discussed in more detail in Appendix A. The demand for air travel is also likely to be influenced by the relative travel time and cost of alternative modes. Some studies have included cost and travel time variables for alternative surface modes. Functional Form The majority of econometric models estimated in academic air travel demand studies have adopted a functional form in which the dependent variable and the continuous independent variables (such as travel time, distance, or cost) are expressed in logarithms. This model form has the advantage that the estimated values of the coefficients of the continuous vari- ables give the elasticity of demand with respect to that variable. However, this implies that the elasticity is constant. A few studies have used more complex model forms, such as the trans- log model, which includes second-order terms (e.g., Oum and Gillen 1983; Oum, Gillen, and Noble 1986). Estimation Data Air travel demand studies have been based on many different data sources. In some cases this is due to the different markets that were studied (e.g., U.S. domestic markets vs. European markets). However, even when the same data source was used, such as the U.S. O&D passenger data from DB1B, different studies used different city-pair markets, different time periods, or different temporal resolution (quarterly vs. annual data). Combined with differences in model functional specification and explanatory variables, this makes it difficult to isolate the underly- ing reason why results differ across studies. Estimation data can be broadly classified as time- series, cross-sectional, or panel, where panel data combines the strengths of cross-sectional and time-series data by tracking multiple markets over time. Of the 45 models used in the academic studies reviewed, 14 used time-series data, 13 used cross-sectional data, and 16 used panel data. One model relied on ticket sample data and another used passenger data from air passenger surveys. Aggregated versus Disaggregated Data The review of the academic literature found relatively few examples of air travel demand studies that used disaggregated data. More recent studies include those by Dargay (2010), who developed a model of person-miles of domestic intercity travel by mode (including air) in the United Kingdom (UK) at the household level for five different trip purposes using data from the UK National Travel Survey from 1995 to 2006, and Kressner and Garrow (2012), who developed a model of the number of home-based air trips generated by households at the zip code level in Atlanta. Two studies examined air travel demand using airline booking data (Granados, et al. 2012; Mumbower, Garrow, and Higgins 2014). Although relatively few of the academic studies reviewed used air passenger survey data, this represents a valuable source of potential information about air passenger characteristics that could improve studies of air travel demand (Gosling 2014). Market Segmentation Only five of the 45 air travel demand models reviewed for the project distinguished between demand for business and non-business purposes. Morrison and Winston (1985) used U.S.

Survey of Past Analyses of Air Passenger Demand 11 National Travel Survey data to estimate separate models for vacation and business travel. Dargay (2010) used data from the UK National Travel Survey to estimate two different models of person-miles of travel within the UK by mode (including air) for five different trip purposes. Granados, et al. (2012) developed separate models for business and leisure travel, using airline booking data. Although factors that determine the level of air travel for business trips are likely to be dif- ferent from those that determine the level of nonbusiness trips, the lack of efforts to assemble reliable data on trip purpose on an on-going basis other than in the UK has limited the develop- ment of air travel demand models that account for differences in demand by trip purpose. Government and Industry Studies The research reviewed a number of government studies of air travel demand that were con- ducted for Australia and the UK and two industry studies that were conducted for the Inter- national Air Transport Association (IATA) and the Airports Council International–North America (ACI-NA). Australian Studies A number of air travel demand studies (domestic and international) were performed by or for Australian government agencies over a 16-year period from 1982 to 1998, reflecting the importance of air travel to Australia. Two early studies used elasticity values for Australian air travel obtained in prior studies to understand the sensitivity of air travel demand to price and other factors. Lubulwa (1986) describes the development of demand functions for long distance travel in Australia by air, rail, long-distance bus, and private car, using elasticity esti- mates from prior studies. The second study, undertaken as part of an independent review of economic regulation of domestic aviation in Australia (May, Butcher, and Mills 1986), ana- lyzed the sensitivity of air travel demand to changes in airfares. Their analysis reported price elasticities for selected Australian domestic routes estimated in a 1985 study by the Bureau of Transport Economics. The same year, Australia. Bureau of Transport Economics (1986) issued a report that pro- vided forecasts of domestic air passenger and air freight demands in Australia. The report reviewed recent trends in passengers and air fares for trunk, regional, and commuter markets. Econometric models were then estimated for each of these markets using quarterly data for 1977 to 1984. The models predicted the number of passengers on nonstop flights using city-pair or regional population as explanatory socioeconomic variables together with GDP (for trunk routes and regional air services) or average male weekly earnings (for commuter air services) as income variables. Other explanatory variables included airfares and measures of the cost of alternative surface travel, both expressed as a price index. A later report by Australia. Bureau of Transport and Communications Economics (1995) documents the development and estimation of econometric models of the demand for air travel between Australia and 12 foreign countries using quarterly data from 1986 to 1993. For each country-pair, separate models based on a double-log or linear functional form were estimated for four market segments: Australian residents traveling for business, Australian residents traveling for leisure, foreign residents traveling for business, and foreign residents traveling for leisure. A subsequent paper by Hamal (1998) presents the results of a regression analysis of Australian resident holiday travel in two long-haul (UK and United States) and four short- haul markets (Fiji, Indonesia, New Zealand, and Singapore) using annual data from 1974 to 1996. The models predicted per-capita resident vacation departures as a function of per-capita real household disposable income, the price index of domestic travel and accommodation, the

12 Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies destination country’s consumer price index (CPI), used as a proxy for a price index of destina- tion travel and accommodation costs, and the annual average exchange rate. United Kingdom Aviation Forecasts Since the 1990s, the UK Department for Transport (DfT) and its predecessor agencies have been developing a sophisticated set of aviation forecasting models. The overall modeling framework reflects policy needs of the DfT to address airport capacity issues in the London region and to develop a national airport policy. The UK aviation forecasting model divides the UK into 455 analysis zones and distributes air trips to and from each zone to the national system of commercial service airports. The most recent report on the national aviation fore- casts (UK DfT 2013) describes the modeling framework and its use to prepare forecasts of commercial aviation activity at all UK airports with significant levels of commercial air service. The framework consists of several linked models that: (1) generate forecasts of air passen- ger traffic for the country as a whole in six different geographic market sectors; (2) allocate the national passenger traffic in each market sector to airports; and (3) estimate the number of air transport movements (aircraft operations) at each airport resulting from the passenger allocation. Separate models were estimated for six geographic market sectors: Western Europe, other Organisation for Economic Co-operation and Development (OECD) countries, newly industrialized countries, less developed countries, UK domestic traffic, and international-to- international connecting traffic using UK airports (primarily London Heathrow). Passenger traffic between the UK and the four overseas market sectors was further stratified into four trip purpose segments: UK residents or foreign residents making business or leisure trips. UK domestic passenger traffic was stratified into business and leisure trips and international con- nections were not stratified by trip purpose. The explanatory socioeconomic variables varied with the market sector, with one or two variables for a given sector, selected from UK GDP, UK consumption, UK imports, UK exports, or foreign GDP. Dummy variables for years in which there was a special event (such as a recession or the 9/11 terrorist attack) that were statistically significant were included in some models. The general form of the econometric demand models used a log-log structure. Lagged as well as “difference” terms (defined as the difference in a vari- able from the previous year) were included in the models. Industry Studies In recent years two consultant studies have been undertaken for industry organizations that have included the development of models of air travel demand (InterVISTAS 2007; InterVISTAS 2014). Although the primary focus of both studies was on air travel demand price elasticities, both studies developed models of air travel demand that included socioeconomic variables. The 2007 study developed nine models: a world model and separate models for eight global regional markets, including the U.S. domestic market, transpacific, and transatlantic markets. The models for the U.S. domestic market in both the 2007 and 2014 studies used essentially the same data and model structure, although for different years. The 2007 study developed a model of O&D passengers for the top 1,000 city-pair markets defined on a metropolitan area basis for the period 1994 to 2005. The 2014 study updated this model using data for the period 2000 to 2010, but limited the analysis to the top 500 city-pair markets. The 2007 study used several explanatory variables (the variables included in the nine models varied). These explanatory variables included airfare, income measures based on GDP, popu- lations of each country-pair, market distance, and quarterly and monthly dummy variables. Explanatory variables used in the 2014 study included the metropolitan area population, real per-capita personal income, real average fare in the market (calculated in different ways),

Survey of Past Analyses of Air Passenger Demand 13 dummy variables for the hub size of airports at each end of the market, and quarter and market dummy variables. ACRP Reports and Other ACRP Documents Several past ACRP reports address airport issues related to airport forecasts of air passenger demand. Most significantly, ACRP Synthesis 2: Airport Aviation Activity Forecasting (2007) reviews the methods and data that have been used to conduct aviation activity forecasting by airports, including passenger activity forecasts. The study covers the elements of aviation demand that an airport may seek to forecast, including aircraft operations and passenger enplanements, the information and data sources that may be used to develop forecasts, the methods available for creating forecasts, and the approaches that may be used for evaluating forecasts. The discussion of the drivers of airport aviation activity mentions the importance of economic and demographic factors, but does not address potential roles for disaggre- gated socioeconomic data. Practitioners will find this report to be a useful companion to the current report. ACRP Report 18: Passenger Air Service Development Techniques (2009) addresses the ways in which airports can assess the passenger air service available for their passengers and reach out to airlines to enhance and expand these services. The report notes that airlines would expect such an airport to provide credible forecasts of potential passenger activity, based on the airport’s history and characteristics, but does not address how these forecasts can be developed. ACRP Synthesis 7: Airport Economic Impact Methods and Models (2008) presents an overview of economic impact modeling. Frequently conducted by airports of all sizes, this type of analysis identifies the scale of regional economic activities that can be associated with the operations of an airport. Although these impacts will be related to the level of passenger activity at an airport, the report does not cover the passenger forecasting process. ACRP Report 26: Guidebook for Conducting Airport User Surveys (2009) provides guidance on designing and conducting airport-user surveys, including surveys of air passengers, airport employees and tenants, area residents and businesses, and collection of air cargo data. Although not addressing air travel demand directly, the data generated by surveys of air passengers, house- holds, and local businesses can be a source of information for air travel demand studies. By its very nature, such data is highly disaggregated, allowing a level of analysis not possible with more aggregate data. ACRP Report 76: Addressing Uncertainty about Future Airport Activity Levels in Airport Deci- sion Making (2012), provides a guidebook to assist airport planning and management personnel in addressing uncertainty about future air traffic levels in making airport development decisions. The report contains a number of examples of airports where air passenger activity has evolved in a different pattern from that projected in successive forecasts over a period of years and describes various techniques that can be used to address the uncertainty inherent in forecasts of air passenger demand and resulting aircraft activity. The report mentions the use of causal models of air passenger demand, but does not discuss specific models. ACRP Report 98: Understanding Airline and Passenger Choice in Multi-Airport Regions (2013) addresses factors that influence airline decisions on what air service to offer at different airports in a multi-airport region and factors that influence air passenger decisions on which airport to use for a given trip depending on air service offered and any fare differences. The report includes five regional case studies that describe the geographic and economic context in each region, the evolution of air service at the airports serving the region, and the resulting air passenger traffic levels. Although the report discusses many of the underlying socioeconomic factors that drive

14 Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies the demand for air travel, it does not address how those interact to influence the total regional air travel demand. The report does not present a formal model for predicting airport choice in a multi-airport region, but it does include a fairly comprehensive literature review that contains summaries of many prior studies that have developed such models. ACRP Synthesis 51: Impacts of Aging Travelers on Airports (2014) addresses the preparations and accommodations that airports can make as the percentage of air travelers who are older increases, reflecting trends in the general population. Although this report does not address passenger forecasting as such, it does raise socioeconomic and demographic issues that over- lap with potential uses of disaggregated socioeconomic data by airports in their forecasting activities. The report provides a detailed study of changing U.S. demographics, in particular the growing size of the older or elderly populations. Such populations have unique characteristics and specific needs, and they are likely to make up an increasing share of the air traveling public. For this reason, airport managers and staff should be aware of the expectations and needs of these airport users, especially in such areas as mobility within airports and the availability of clear information for these airport users. Although the report contains demographic information about older travelers, there is no discussion of their demand for air travel. ACRP Web-Only Document 22: Passenger Value of Time, Benefit-Cost Analysis, and Airport Capital Investment Decisions (2015) reports the findings of ACRP Project 03-19. A web-based survey of the most recent air trip undertaken by survey respondents within the preceding six months was part of the project. Although the primary purpose of this survey was to perform stated-preference choice experiments in order to estimate differences in the perceived values of time for different components of the air trip by the survey respondents, the survey also collected detailed information on the most recent air trip, including the trip purpose, and information on household characteristics of the survey respondents. ACRP Report 132: The Role of U.S. Airports in the National Economy (2015) includes an analy- sis of the effect of changes in air service on business productivity by different sectors of the economy and an analysis of the effect of changes in airfare on consumer surplus and hence macroeconomic indicators. Although the project did not develop explicit models of air passen- ger demand, it included a review of the literature on air travel demand modeling to identify prior estimates of price elasticity for non-business travelers. It also assembled a database of prior air passenger surveys to estimate the proportion of non-business travel in a sample of 100 domestic airport-pair markets. ACRP Report 142: Effects of Airline Industry Changes on Small- and Non-Hub Airports (2015) identifies airport policy and planning options that provide a response to recent changes in the airline service patterns and offerings at smaller airports. It offers managers of these airports background resources and information for planning for and responding to changes in airline services. Although the report does not explicitly address passenger forecasting and the use of disaggregated socioeconomic data for passenger forecasting, improving an airport’s ability to forecast future passenger demand may be able to contribute to those small airports’ responses to a changing air transportation industry. One of the ACRP reports reviewed notes the importance of accounting for economic and demographic factors in forecasts of air travel demand when addressing service expansion, but does not address potential roles for disaggregated socioeconomic data. Other ACRP reports provide valuable guidance for airports in specific areas, such as how their facilities and opera- tions can be adapted to better accommodate the mobility needs of an elderly population. These reports also provide disaggregated datasets that can be leveraged for the current project to better understand the role of disaggregated socioeconomic factors in producing airport activity forecasts.

Survey of Past Analyses of Air Passenger Demand 15 Surveys and Studies of Traveler Behavior and Choice Analysis of traveler behavior underlies all attempts to model air passenger demand, and more generally intercity travel demand, although often this analysis is implied by the form of the model used. Air passenger traffic at a given airport is the result of several choices made by poten- tial travelers, including the choice of whether to make a trip at all, which mode to use, and if air travel is chosen, which airports to use. Other travel choices (in the case of air trips) include how to travel to and from the O&D airports used and the airlines and flights chosen. Although these choices are often viewed as separate or dependent decisions once the basic decision to make an air trip has been taken, in reality one cannot separate these decisions from the overall level of air travel demand. This is most apparent in the case of air passenger traffic at smaller airports or secondary airports in metropolitan regions, where airport choice directly determines the level of traffic. At many smaller airports some air travelers make the choice of using surface modes to travel to or from a more distant airport where better or less expensive air service is available (an issue sometimes referred to as “leakage”). Intercity travel demand studies were included in the literature review for three reasons. The first is that many of these models generate projections of air travel in addition to travel demand for surface modes. Indeed, although many air travel demand studies (in fact the great majority of such studies) do not consider the effect of competition from other modes, in reality competi- tion from surface modes, particularly personal vehicles, can be significant for trips up to about 1,500 miles (U.S. Bureau of Transportation Statistics 2006). Furthermore, past studies have shown that in countries that have made a major investment in high-speed rail systems, many users of those services would otherwise have flown to their destination, as evidenced by the drop in air travel between London and Paris or Brussels with the opening of the Channel Tunnel and completion of the high-speed rail link between the two cities (Behrens and Pels 2012). Therefore intercity travel demand studies that generate projections of air travel are one type of air travel demand study, although they may not be thought of as such, and the models developed in these studies may shed some light on the role of socioeconomic factors on air travel demand. A second reason for including these studies is that, in contrast to the majority of air passenger demand studies described in this review, which have focused on explaining past levels of air travel, many intercity travel demand studies have been undertaken with the explicit purpose of generating projections of future travel demand, including air travel demand. This forward- looking aspect is likely to be of particular relevance for airport practitioners and planners. The third reason is that a number of these studies have made use of disaggregated socioeconomic data. In contrast, almost all the studies of air travel demand identified in the literature review have used aggregated measures of socioeconomic variables. Intercity travel demand studies typically examine the choices made by travelers for trips between cities with respect to travel mode and may also consider the number of trips made. These trips, often referred to as long distance trips, include nearly all commercial air trips. These studies typically construct models of traveler choice between the travel alternatives faced by these travelers that express these travel choices as probabilistic functions of traveler characteristics and modal features. Data on intercity travel mode choices in the United States from national household travel surveys show that for trips of up to about 750 miles each way a higher propor- tion of travelers use surface modes, particularly personal vehicles, than use air travel (Figure 1). Even for one-way trips between 750 and 1,500 miles personal vehicles account for a significant share of all travel. This suggests that changes in the relative costs, travel times, and convenience of air travel and surface modes are likely to have a significant effect on air travel demand, at least in markets below about 1,500 miles. The choice of mode for a trip of a given distance is also likely to be affected by the socioeconomic characteristics of the travelers, which will influence how they trade off the different attributes of each mode.

16 Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies There is an extensive literature on intercity travel demand that shows how analysis techniques have evolved since some of the early abstract mode work on the U.S. Northeast Corridor in the 1960s (e.g., Baumol and Quandt 1966). Two NCHRP reports (Horowitz 2006; Cambridge Systematics 2008) provide an overview of the evolution of these analysis techniques. A number of more recent studies discussed in more detail in Appendix A document subsequent develop- ments and the current state of practice. Several of these studies developed or applied intercity node choice models. Baik, et al. (2008) developed a mode choice logit model to allocate intercity trips across two existing transporta- tion modes, including commercial air travel, and one hypothetical mode. Cambridge System- atics (2014a and b) presented results from a long distance travel demand model developed to estimate ridership on proposed California High Speed Rail (HSR) passenger services. Outwater, et al. (2015) summarized a multi-year research project using a tour-based simulation framework to model long distance passenger travel by U.S. households involving trips by air, rail, bus, and automobile. Air Passenger Demand Studies by Airports Air passenger demand and passenger enplanements are, not surprisingly, important elements in many types of analyses of airport performance and prospects. For both internal and exter- nal audiences, these airport planning analyses naturally involve the preparation of forecasts of passenger demand and activity at the airport. This section examines how a variety of airports have used socioeconomic data in their air passenger forecasting efforts for use in these varied analytic contexts and purposes. The following examples are presented in much greater detail in Appendix B. Master Plan Studies An airport master plan is an analytic tool or document for determining the long term develop- ment plans for an airport. It includes planning elements from every facet of airport operations, all of which have an objective of contributing to the airport’s ability to handle future passenger One-way trip length (miles) P er ce n ta g es Figure 1. Mode use by trip distance. Source: U.S. Bureau of Transportation Statistics, 2006 (Figure 2).

Survey of Past Analyses of Air Passenger Demand 17 demand as a shared performance requirement or objective. Airport master plans nearly always include forecasts or projections of air passenger activity, and they are a natural place to look for examples of the ways airports and their consultants model and predict passenger activities such as annual enplanements. An important part of the baseline research for this project was to analyze and assess the modeling approaches used by and for airports to model air passenger demand and air passenger activity. Similar studies that are strongly driven by an underlying model of air passenger activity at individual airports are airport environmental or noise assessments and other forms of demand study. Of the 47 studies of air passenger demand conducted by airports or by consultants spon- sored by airports that were reviewed as part of this research (master plans, demand studies, and environmental or noise assessments) a majority acknowledge the role of socioeconomic data in determining air passenger activity, although socioeconomic factors were included in the air passenger demand forecasting models developed in the studies less often. Three of the docu- ments did not provide enough detail to identify how or whether the airport used socioeconomic data in its analysis, but of the 44 that did refer to a use of socioeconomic data, 39 also contained information about the forecasting approaches that were used. Only 21 of the 39 reported using socioeconomic data in their forecasting calculations. Some of the airports use forecasts from the FAA’s Terminal Area Forecast (TAF) to project their passenger demand, while others develop their own analyses and forecasts based on the spe- cific conditions present at their airport and in the surrounding region, including socioeconomic factors. For these forecasts, airports or their consultants tend to use one of three broad types of analysis. The first, time series or trend analysis, reviews the history of an airport’s passenger activity and bases its passenger enplanements forecast on that history, under the assumption that in the absence of major industry changes, observed trends will continue into the future. A second approach, market share analysis, reviews forecasts for the U.S. aviation industry as a whole, or for the region containing the airport, and develops predictions for passenger activity using the assumption that the airport’s share of that overall aviation market will remain as it had been, or will change in a well-defined way. These two types of analysis differ in that time series analysis extrapolates past trends, whereas market share analysis may allow for adjustments that can reflect the effect on the airport’s market share of changes in the aviation industry, such as mergers, creation of new flight services, or moves towards certain types of flights. The third type of analysis, econometric or regression modeling, uses statistical methods to estimate the historical relationship between the airport’s passenger enplanements and sets of selected independent variables, such as national variables, local or regional socioeconomic vari- ables such as population, economic activity (GDP), or income. These models typically use linear or logarithmic regression techniques. Whereas the first two methods may indirectly take into account socioeconomic changes (e.g., an increased preference for low-cost carriers nationwide would increase flight forecasts for those carriers across the country, which would be reflected in market share analyses at airports used by these low-cost carriers), econometric models use a more explicit treatment of the links between socioeconomic variables and the airport’s over- all passenger demand as represented by passenger enplanements. The regression approach also allows the analysis to examine the significance (in scale and precision) of the relationship between air passenger demand at the airport and the individual independent variables. Of the airports that reported they used regional socioeconomic variables in their forecasting efforts, the majority stated that population, income or economic activity, and/or employment were relevant for understanding and forecasting air passenger traffic. To calculate forecasts of passenger enplanements with these models, forecasts of the independent variables used in the models must also be available to the analyst. Such projections for socioeconomic variables are commercially available (from firms such as Woods & Poole Economics, Inc.) as part of the

18 Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies data products prepared for use by entities and organizations in academia, government, and the private sector (including airports) that conduct quantitative economic research that relies on relationships between socioeconomic variables. Among the airports examined in our research, regional population, regional GDP or GRP, and regional income are frequently used independent socioeconomic variables for statistical esti- mations. Among those airports, regional variables were often more statistically significant than employment levels for individual airports, although airports created models using different geo- graphical areas [e.g., city, metropolitan statistical area (MSA), and state level socioeconomic data and projections] and often defined variables differently. Other socioeconomic variables mentioned or used in some of the airport master plan forecasts included cost of living, employment by sector, and, for areas with economies that were particularly reliant on tourism, variables such as de facto population due to tourism or the prevalence of second homes in the region served by an airport. The analysis of the sample of airport master plan forecasts is presented in Appendix B. The airports’ use of forecast modeling techniques or approaches for air passenger demand is sum- marized in Table 1. In the table, trend analysis is time series and all types of regression methods (linear and logarithmic) are categorized together as regressions. One non-hub airport created forecasts of passenger enplanements by predicting a future annual growth rate in passengers. This is similar to the time series method, but because this airport used other information to determine the preferred growth rate, it has been categorized separately. Ten airports used mul- tiple models to develop passenger activity forecasts, while seven relied on the FAA’s Terminal Area Forecast. Fifteen airports used a regression model of some kind to model air passenger demand and estimate forecasts for annual enplanements at their airport. Other Types of Airport Studies Examples of other types of air service studies or airport studies that use forecasts of airport passenger activity or rely on socioeconomic variables in some way were studied. These examples are analyzed in Appendix B, and are briefly summarized here. Because the principal business or activity of most commercial service airports is accommodating demand for flight services and air travel, it is not surprising to find interest in modeling and predicting passenger activity present in a variety of types of airport studies. FAA Terminal Area Forecast The FAA prepares forecasts of aviation activity at each active airport in the National Plan of Integrated Airport Systems (NPIAS) which are published annually as the FAA Terminal Area Preferred Model Large Medium Small Non-Hub Total Growth Rate - - - 1 1 Market Share 1 2 2 1 6 Regression 7 4 3 1 15 TAF 3 1 1 2 7 Time Series - - - 2 2 Multiple Models 3 3 3 1 10 Not Specified/Unknown 1 - 3 2 6 Total 15 10 12 10 47 Table 1. Passenger air travel demand models used in sample of airport forecasts.

Survey of Past Analyses of Air Passenger Demand 19 Forecast, or TAF. As of the end of 2017, the most recent TAF provides forecasts for the period 2016 to 2045 (FAA 2017). The TAF provides historical data and forecast activity levels on a Federal fiscal year basis for the following measures of aviation activity for each airport: • Enplanements (enplaned passengers) by air carriers and regional airlines (separately and total) • Itinerant aircraft operations by air carriers, commuters and air taxis, general aviation, and military aircraft (separately and total) • Local aircraft operations by civil (general aviation) and military aircraft (separately and total) • Total aircraft operations • Based aircraft In addition the TAF includes historical data and forecast activity for total aircraft operations under radar control at each Terminal Radar Approach Control (TRACON) facility. With effect from the 2014 TAF, a new forecast process was adopted, termed the Terminal Area Forecast Modernization (TAF-M) (Bhadra 2013 and 2014). This is based on a set of formal, bottom-up models of air travel demand and aircraft operations (LeBoff 2016). The new forecast process represents an ambitious attempt to develop a comprehensive air travel and aircraft activity forecasting process for the United States at a national level. Among the aspects of the process is a consistent forecasting approach for all airports in the United States that is based on an underlying model of air travel demand that reflects the influence of socioeconomic factors (currently limited to total personal income) as well as differences in air service in each O&D market (currently incorporated as average airfares in the market and the number of avail- able routes serving the market). An important benefit of the bottom-up approach to modeling O&D passenger demand is that the resulting forecasts of enplaned passengers and commercial aircraft operations can be made on a flight segment basis, rather than simply on the basis of total activity at a given airport. This new degree of specificity provides benefits for use of the resulting forecasts in a wide range of planning activities, from aircraft noise analysis at airports to airspace and air traffic management planning, and is a significant improvement over what was essentially trend analysis of airport activity. However, there are many aspects that appear deserving of continued research and improve- ment. Because the TAF has to cover a very large number of airports (essentially all those in the NPIAS) and needs to be updated on an annual basis, a considerable effort has been devoted to the extensive data management issues involved. Now that the data management challenges have been largely overcome, the influence of different socioeconomic factors on air travel demand and how the influence of these factors may vary in different markets can be addressed. State Aviation and Regional Airport System Plans Five state aviation system plans (Alabama, Florida, Kansas, Minnesota, and Washington) were reviewed; two used the TAF forecasts directly and one primarily adopted airport master plan forecasts, supplemented with the TAF forecast for one airport that presumably did not have a master plan forecast that could be used. The system plan for Minnesota developed forecasts for all the commercial service airports in the state. These forecasts were generated using an econometric demand model with only one socioeconomic variable, the aggregate total personal income in the catchment area of each airport. The system plan for Washington state used a combination of approaches for forecasts for different airports in the state. For Seattle-Tacoma International Airport, the principal airport in Washington (as of the date of this report the system plan accounted for about 87% of the total enplanements in the state), the forecast was adopted from the TAF. For the next eight largest airports in terms of passenger enplanements, which account for a further 13% of total enplanements in the state, forecasts were prepared using three different methods. The average of the forecasts generated by each method was used for each airport, although the details of the three methods are not described in the technical report

20 Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies documenting the forecasts. The forecasts for the remaining airports, which account for less than 1% of the enplanements in the state, were basically assumptions, either by the study team or based on discussions with the air taxi operators serving the airports or the airport management. Four of the five system plans included a discussion of socioeconomic trends in the state, typically including population, employment, and personal income, although only one of the system plans actually used these data in developing the forecasts. Although the socioeconomic data often was presented at the county or regional level, it was generally not disaggregated in any way other than geographically with the exception of the Florida Aviation System Plan, which presented data on historic trends and projections of population by age range and regional dif- ferences in the growth rate of population due to in-migration versus natural increase, although these data were not apparently used in developing the forecast. Airport system studies for five regions also were examined: the New England, Baltimore- Washington, New York/New Jersey, the San Francisco Bay Area, and southern California. The fore- cast approach followed in the regional system planning studies varied in the approach to modeling air travel demand and in the level of detail the modeling was performed. Two different studies for the New York/New Jersey region that used different approaches were reviewed. The studies for all five regions (only the earlier of two studies in the case of the New York/New Jersey region) developed models or procedures to allocate regional or airport O&D demand to regional analysis zones. The resulting distribution of ground trip ends was then used in most studies to assign forecast regional demand to individual airports under different air service scenarios or to forecast airport ground access trip patterns. Two studies, the earlier study for the New York/New Jersey region and the study for the San Francisco Bay Area, estimated trip generation models, while the other studies used a demand allocation procedure based on population, households, employment, and/or income. Based on these regional studies, there appears to be no consistent approach to preparing air passenger forecasts for regional airport system planning studies, although generally some form of econometric model is used, at least for domestic O&D passengers. These models typically use one or two aggregate socioeconomic variables as well as average airline yields. There is no con- sistency in whether the models predict passenger traffic at the market, airport, or regional level, or how the ground origins or destinations of the air trips given by the resulting regional forecasts are allocated to travel analysis zones in the region. Airport Planning Studies Outside the United States Six international demand forecasting studies were examined in the review of international practices. These studies used a variety of analysis methodologies, and they also varied in the extent to which the reports described methodological details. The six examples include studies of the London region by the United Kingdom Airports Commission (2015), and individual air- port studies by the Geneva, Switzerland, Airport Authority (2014), the Dublin Airport Authority (2006), the Airport Authority Hong Kong (2011), the Sydney Airport Master Plan 2033 (2014), and the Greater Toronto Airport Authority (2007). As best can be determined from the documentation for each study, all but one used a multi- plicative (log-linear) model structure, although two studies (for the London region and Canada) used a more complex variation on a traditional log-linear formulation. The study for the London region used lagged and difference terms (described as an unrestricted error correction model) while that for Canada used a Box-Cox model structure. All six studies used aggregated values of GDP as the primary (or only) socioeconomic variable. The forecasting model for the UK (which formed the basis for the forecasts for the London region) also used consumption, imports, and exports as socioeconomic variables in different market sectors, for some market sectors in con- junction with GDP, and for other market sectors in place of GDP. However, the decision of

Survey of Past Analyses of Air Passenger Demand 21 which variables to use for a given sector appears to have been based on which variables gave the best statistical fit rather than any underlying causal logic. Air Service Development and Passenger Leakage Studies Air service development studies and passenger leakage studies are customarily products pre- pared for airports by aviation consulting firms. Air service development studies may be used by airports for marketing efforts about individual travel markets (rather than an airport’s overall passenger demand), which are directed at airlines, often specific airlines. In contrast, passenger leakage studies are generally informational studies for airport stakeholders, with the potential to be used as background for subsequent air service development efforts. For these reasons these studies are often regarded and treated by consultants and their airport clients as proprietary products rather than public documents. Airport air service development studies and the factors customarily considered in them have been described in detail in ACRP Report 18: Passenger Air Service Development Techniques (2009). Along with financial, marketing, and route structure considerations that may be included in an airport’s efforts to develop new passenger services with airlines, the study stresses the value of passenger demand projections and regional demographic and economic factors in influencing decision makers at potential airline service providers. The report identifies the types of socioeconomic data that could inform these studies, but does not propose specific modeling approaches or formulations for use in preparing passenger demand projections. The importance of demographic data for air service development analyses was recently assessed for an audience of airport practitioners (Dietz 2014). Demographic data about the characteristics of the population served by an airport provides airlines with impor- tant information about the likelihood that airline service could be successful. Important factors affecting air service development assessments in these data are the size and characteristics of the airport region’s overall population, the characteristics of the region’s business community, and the characteristics of the region’s potential leisure travel and travelers. ACRP Report 18 also identifies the competitive challenges that are posed by larger airports within driving distance as the principal cause of passenger leakage from a given airport. Such airports are likely to be able to provide business travelers with a wider range of destinations, often at lower fare levels. Lower fare levels will also be attractive for leisure travelers, who are more price sensitive than business travelers in most instances. Airport Choice and Airport Ground Access Mode Choice Studies Although airport demand allocation and airport ground access models are not air travel demand models in the sense that they predict how the overall demand for air travel is influenced by changes in socioeconomic and air service factors, they typically make use of disaggregated socioeconomic data. By predicting how air travel demand in a region distributes itself among the airports serving the region, they can shed light on the level of air travel activity that occurs at a given airport. While airport ground access mode choice models are designed to predict the use of different access modes at a given airport, and do not directly attempt to model the level of air travel activity at that airport, they can form an important component of airport demand allocation models. Since air travelers’ choice of airport in a region served by multiple airports is influenced by the relative accessibility of each airport (Parrella et al. 2013), which depends on the different ground transportation models available for travel to each airport and the relative level of service of each mode, including travel times, frequency, costs, reliability, and number of changes of mode or line required, airport ground access mode choice models can be used to help define the overall accessibility of a given airport from a given location. As such, they will often form a lower-level nest in a nested model of air traveler airport choice.

22 Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies Both airport choice and ground access mode choice models are inherently disaggregate in approach, generally modeling the airport choice or mode choice decisions of individual air par- ties. These individual choices will reflect the different circumstances faced by each air party, including trip ground origin, air trip destination, time of day of travel, and travel party charac- teristics, all of which influence each air party’s decisions. Typically these studies use data from an air passenger survey to define the travel and air party characteristics of a reasonably large sample of air travelers to estimate disaggregate behavioral choice models, which can then be applied to develop forecasts of future airport use patterns or ground access travel. The extent to which these models include demographic or socioeconomic factors varies, often depending on the data available from the air passenger survey from which they are estimated. Airport Ground Access Mode Choice Models The state of practice of airport ground access mode choice models was reviewed in an early ACRP Synthesis study (Gosling 2008). This study included a detailed review of nine airport ground access mode choice models. Five of the nine models included household income as an explana- tory variable, although the way that this was done varied across the five studies. A model to analyze the potential diversion of air trips to intercity rail from improved rail service that was developed as part of a recent ACRP research study on integrating aviation and passenger rail planning (Resource Systems Group, Inc. et al. 2015) included a sub-model of airport choice or station choice. In another study, the joint airport choice and ground access mode models developed in the New York Regional Air Service Demand Study (RASDS) and the West of Hudson Regional Transit Access Study (WHRTAS) incorporated a limited number of socioeconomic variables in the access mode utility functions. The RASDS models included gender, age, and income, although age and income were only expressed in three ranges. The WHRTAS models included the average number of automobiles per household member, but only in the utility function for the drive and park mode for trips by residents of the region and used the average value for the analysis zone of the trip origin (presumably because the data on household automobile ownership and household size were not available from the air passenger survey, although this was not explained in the report). There has been limited experience using socioeconomic variables in airport ground access mode choice models. Where these have been included, they have been limited to gender, age, household income, household size, and automobile ownership (number of automobiles per household). Only one model identified in the review of relevant literature included gender or age, with the age variable based on three age ranges. Only one other model included household size, and this was limited to a dummy variable for single-person households. Two models con- sidered household automobile ownership, with one using the average ratio of automobiles to household members for the analysis zone of the trip origin and the other using a dummy variable for households with fewer automobiles than employed household members. Household income was considered in several models, but the way in which income was incorporated in the models varied widely from dummy variables for incomes in a specified range to continuous functions of income. Thus while it appears from the findings of past studies that household income, and possibly gender and age, influence airport ground access mode choice, there is no consensus yet on how best to incorporate these factors into airport access mode choice models. Airport Bond Prospectus Documents Six airport bond prospectus documents were reviewed and all include some mention of socio- economic factors as determinants of air travel demand that can be used to develop forecasts or projections of future air passenger enplanements. While none of the documents provide infor- mation on air travel demand model structure or parameter estimates, in several cases the fore- casts were made using a regression model that includes some aggregate socioeconomic factors as

Survey of Past Analyses of Air Passenger Demand 23 independent variables. Some of the documents rely on more than one forecasting methodology to characterize the factors that affect air passenger demand at the airport. Of special interest for this study, three of the studies treat region-specific disaggregated socioeconomic factors related to income distribution, age distribution, and ethnicity or national origin as contributors to the regional economic basis for air travel demand at the airport, but these factors are not included in air passenger demand modeling portions of the reports. In some cases the bond documents use arguments about the contribution of disaggregated socioeconomic community features in qualitative rather than quantitative or computational ways. Passenger Demand Studies by Airlines and Other Industry Stakeholders This section examines interest in passenger demand modeling and forecasting for other com- mercial aviation industry stakeholders such as airlines and commercial jet manufacturers. Not surprisingly, airlines pay greater interest to details that distinguish among individual travel markets than manufacturers, who take a much broader perspective on passenger demand issues. An overview of one airline’s approach to using socioeconomic data to better understand and serve passenger demand was presented at the TRB Annual Meeting in 2014 by Andrew Watterson of Southwest Airlines (Watterson 2014). The airline objective for using these data was described as “better understand[ing] passengers and best match[ing] [Southwest’s] supply to their demand.” With regard to these passengers, the airline was interested in these questions: • Why they fly? • Where they fly? • Can they afford to fly? • Which airport do they prefer? The presentation identified several sources for socioeconomic data, including federal sources [Census Bureau, Bureau of Labor Statistics (BLS), Bureau of Economic Analysis (BEA), Bureau of Transportation Statistics (BTS)]; survey and research firms providing data that was purchased by the airline; local sources of data such as community and regional convention and visitors’ bureaus; research sponsored by airports served by the airline; and internal airline data on pas- senger behaviors, preferences, and destinations. A number of aircraft manufacturers, airline industry associations, and other industry organi- zations prepare forecasts of future growth in air travel demand, primarily to project the expected future demand for their products or services. These forecasts are typically updated on an annual basis. Although these forecasts do not address air travel demand at the level of individual air- ports, the forecast growth in air travel demand at a world regional level projected in these studies is often cited in forecasts prepared for specific airports or airport systems as a “reality check” on the growth in demand projected in those forecasts. Because of the global or national nature of industry forecasts, they generally forecast air travel demand in terms of revenue passenger-miles (RPM) or revenue passenger-kilometers (RPK) and use aggregate measures of economic activity, such as national GDP. The forecast reports may discuss more disaggregate factors affecting air travel demand, but from the limited details on the forecast methodology that are typically reported, it is unclear how these factors are incor- porated into the forecasts. Summary and Conclusions The preceding examples illustrate the wide range of approaches for modeling and forecast- ing air passenger demand, and the wide range of uses for such models and forecasts that exist across the airport and air transportation community. At the same time, the estimated demand

24 Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies and forecast models are nearly always a means of calculating passenger enplanement forecasts to justify or calibrate airport plans and financing instruments. These models usually represent a relatively small portion of the overall airport analysis, especially in the case of a master planning exercise. For these reasons, the statistical parameters for the models underlying air passenger demand forecasts, especially as used in airport master plans and other airport studies of enplane- ments, are rarely reported in any detail. This relative silence regarding the numerical and statistical details of the estimated models for passenger enplanements is understandable because those models serve as stepping stones to the passenger demand forecasts that underlie a wide range of airport planning concerns. Since most passenger demand modeling is done by airport consultants, there may also be reason- able proprietary concerns about revealing too much of a firm’s “analytic toolkit” to the eyes of competitors. However, greater transparency in the reporting of model parameters and other statistical details could encourage a growing and improving base of common industry knowl- edge about passenger demand modeling at the airport level, which could serve the overall airport community well.

Next: Chapter 3 Sources of Disaggregated Socioeconomic Data »
Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s Airport Cooperative Research Program (ACRP) Research Report: 194: Using Disaggregated Socioeconomic Data in Air Passenger Demand Studies explores the potential benefits of using disaggregated socioeconomic data, such as regional household income distributions and air passenger and travel survey data, for air passenger demand studies.

Aviation demand is strongly correlated to socioeconomic activity, and analysts typically use aggregate socioeconomic data, such as gross regional product or average regional household income, to better understand current and potential future aviation demand.

Since the United States is experiencing significant and ongoing demographic trends there is a question as to whether traditional methods and data sources will adequately capture these trends or would more detailed, disaggregated socioeconomic data, or even nontraditional data provide more accurate results.

This report summarizes long-term socioeconomic trends, attempts to understand their potential impact, and provides guidance for incorporating disaggregated socioeconomic data into air passenger demand studies.

The following appendices to ACRP Research Report 194 are available online:

Appendix A: Detailed Survey of Past Analyses of Air Passenger Demand

Appendix B: Airport Industry Use of Socioeconomic Data for Air Passenger Demand Studies

Appendix C: Additional Material on Sources of Disaggregated Socioeconomic Data

Appendix D: Detailed Case Study Analysis Results

Appendix E: Background on Other Analytic Approaches

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!