Chapter 1 Results of National Household Travel Survey

Introduction

The terrorist attacks in the United States in September 2001 had an immediate and visible impact on transportation. While the obvious impacts were temporary, there may have been less obvious yet longer lasting changes in U.S. travel patterns. The purpose of this study is to provide a greater understanding of the passenger travel behavior patterns of persons making longdistance trips before and after 9/11.

The 2001 National Household Travel Survey (NHTS) provides an opportunity to explore some of those changes in travel patterns. The survey was conducted between March 2001 and May 2002. The tragic events of 9/11 happened to occur during the data collection period. As such, even though the survey was not designed to capture 9/11 impacts, it can be used to look at some of the changes that occurred with travel patterns during that period. To examine the changes in travel patterns, the dataset was divided up into a pre-9/11 dataset and a post-9/11 dataset. These datasets were each reweighted by the U.S. population to produce nationally representative annual datasets. These datasets were then analyzed and the results compared to determine what, if any, interesting changes in travel patterns could be identified.

However, there are two major difficulties in using these datasets to estimate the impacts of 9/11 on travel patterns. First, the datasets are not seasonally adjusted. Because the two datasets cover trips for only a portion of the year, seasonal travel patterns can affect the number of trips and trip characteristics within each dataset. Because NHTS data are not collected on an annual basis, there is no way to use previous survey data to adjust for those seasonal patterns. The second difficulty is the changing economic environment during the time the survey was conducted. These changing economic conditions also

likely affected the number of trips and trip characteristics when comparing the pre-9/11 and post-9/11 datasets. As a result of these limitations in the NHTS data, this report also contains two additional 9/11 analyses, a time series analysis using seasonally adjusted historical and forecast data by mode, and an econometric analysis looking at the impact by mode of 9/11 and several economic variables.

Pre- and Post- 9/11 Environment

As discussed above, economic changes can have positive or negative effects on travel, making it hard to isolate the impacts of 9/11. And different economic factors can simultaneously influence changes in opposite directions for different types of travel, as can be surmised from the following discussion.

Prior to September 2001, the economy was experiencing flat or slightly declining growth as real Gross Domestic Product (GDP) remained around the $9.87 trillion to $9.90 trillion range (in chained real 2000 dollars) for over five quarters, from the second quarter of 2000 through the third quarter of 2001. An official recession started from the economic peak in March 2001 until the economic trough in November 2001, according to the National Bureau of Economic Research (NBER). ¹ After 9/11, from the fourth quarter of 2001 to the second quarter of 2003, GDP showed a steady growth

(figure 1). From these data, one might predict an increase in travel after 9/11, commensurate with the recovering economy. Travel did start to recover after 9/11, but it did not quickly return to previous levels. For example, the number of airline passengers did not surpass its 2001 peak until July 2004.

Besides the economic downturn identified by the NBER, there were other signs that the economy had started to weaken prior to 9/11. Industrial production is a good indicator of business activity and reflects the level of business travel and overall production of the economy. The economic downturn is reflected in the rapid decline of the Industrial Production Index (figure 2), which had fallen from a high of 115 in July 2000 to 110.6 by September 2001 as business inventories began to accumulate. Therefore, declining industrial production prior to 9/11 may have contributed to an overall drop in travel. However, after January 2002, the index began to rise and, although it did not reach the levels seen in 2000, it did indicate an economic upturn, which might have contributed to a predicted increase in travel after 9/11.

Beginning in January 2000, the unemployment rate steadily increased from just below 4 percent to almost 5 percent by September 2001, and by May of 2003 had risen above 6 percent (figure 3). It might reasonably be expected that rising unemployment would lead to reductions in personal travel throughout the period covered by the 2001 NHTS.

Even though the air travel price index (ATPI) (figure 4) indicates that real ticket prices were already declining from the first quarter of 2001 through the third quarter of 2001, travel had just begun to respond to lower ticket prices prior to 9/11. The transportation service passenger index (figure 5), a measure of transportation activity strongly influenced by airline travel, was rising in July and August of 2001, from 101.2 to 103.2. After 9/11 the index fell significantly to

79.3. After 9/11, ticket prices, as measured by the ATPI, continued to fall almost to 1995 price levels in the fourth quarter of 2001 and rose slightly thereafter with some volatility, but has not reached the levels of the first two quarters of 2001. With these relatively lower ticket prices after 9/11, the transportation service passenger index began to increase slowly (with a small dip in March 2003 corresponding with the beginning of the Iraq war), but has not reached the levels prior to 9/11. Therefore, it appears that travel was influenced by factors other than just changes in airline ticket prices. Some of these factors will be discussed in the econometric analysis in chapter 3.

NHTS Background

This section of the study examines results from the 2001 National Household Travel Survey (NHTS). NHTS 2001 is a national household survey of both daily and longdistance travel, providing the most recent comprehensive look at travel by Americans. The NHTS data collection occurred from March 2001 to May 2002. The long-distance components of the NHTS 2001 data were divided and reweighted into two nationally representative annual datasets to compare pre-and post-9/11 travel patterns because 9/11 occurred during the NHTS data collection.

Changes in trip volumes, mode choice, and the characteristics of individuals traveling in the United States before and after 9/11 are examined to see if any general patterns can be noted. Also, changes in the proportion of long-distance trips made domestically versus internationally are investigated. Some of the topics covered by mode choice for both the pre-and post-datasets include trip purpose, age, gender, income, trip distance, and trip location.

Long-Distance Travel

Long-distance trips in the 2001 NHTS are defined as trips of 50 miles or more from home to the farthest destination traveled. A long-distance trip includes the outbound portion of the trip to reach the farthest destination as well as the return trip home and any stops made along the way to change transportation modes or for an overnight stay. Long-distance travel includes trips made by all modes, including privately owned vehicle (POV), airplane, bus, train, and ship; and for all purposes, such as commuting, business, pleasure, and personal and family business. Train and ship long-distance travel are combined into an “other” category due to the small number of observations.

NHTS Pre- and Post-9/11 Comparisons²

The annualized full NHTS 2001 data indicate that there were more than 2.6 billion long-distance trips taken in 2001. Approximately 90 percent were by privately owned vehicle (POV). Trips by airplane accounted for 7 percent of long-distance trips. Travel by bus accounted for 2 percent of these trips, and train trips represented less than 1 percent.

The total estimated number of yearly trips was lower using the post-9/11 sample as compared to the pre-9/11 sample – 2.7 billion trips from the pre-9/11 file and 2.4 billion trips from the post-9/11 file, an 11 percent decline. The POV trips were estimated to be 2.4 billion trips from the pre-9/11 file and 2.2 billion trips from the post-9/11 file, an 8 percent decline. Air trips were estimated to be 216 million from the pre-9/11 file and 169 million from the post-9/11 file, the largest decline of all the modes at nearly 22 percent. Further analysis of the two datasets to identify changes in long-distance travel patterns before and after provides additional insight into the possible effects of 9/11.

International Travel Down

Prior to 9/11, international travel (originating from the United States) represented 3 percent of all long-distance trips, but after 9/11 that travel dropped to 2 percent of all trips, a statistically significant decline. The mode with the largest decrease in international travel was air, which fell from 16 to 12 percent of all air trips after 9/11. Air travel accounts for approximately one-half of all international trips. POV long-distance trips for international travel experienced a small decline after 9/11 (figure 6). POV travel accounts for slightly more than 40 percent of international trips.

Personal Business Travel Decreased

Long-distance trips for personal business also experienced a statistically significant decline after 9/11, from nearly 14 percent of all trips to 12 percent. Personal business trips include medical visits, shopping trips, and trips to attend weddings and funerals. POV accounts for approximately 90 percent of all longdistance personal business trips. Air travel, which accounts for only about 5 percent of all personal business trips, experienced a 3 percent decrease from 9 to 6 percent of air trips (figure 7).

More Shorter Distance Trips Taken

The only trip distance that showed a significant increase after 9/11 was in the shortest distance category of between 50 to 99 miles. All modes combined in this distance category showed an increase from 45 to 50 percent of total trips after 9/11 (figure 8). POV trips, which account for more than 95 percent of all trips between 50 to 99 miles, increased from 49 to 54 percent after 9/11, bus trips increased from 23 to 36 percent, and the “other” category increased from 44 to 65 percent. Air was the only mode not to register a significant increase, but air travel accounts for a negligible portion (nearly zero percent) of the trips in the 50 to 99 mile range. Modes that were not statistically significant were not included in the figures.

Appendix A describes the methodology of the NHTS and how the pre-9/11 and post-9/11 data files were created. Appendix B contains tables with the data estimates used in this report and their standard errors. Standard errors are in the same metric as the estimates. All comparisons in the text and graphics are statistically significant at a 0.05 level unless otherwise noted.

Expected Results Not Observed

Immediately following the 9/11 terrorist attack, experts were trying to determine how these attacks would impact our economy, business and leisure travel, and the attitudes of Americans. The terrorist attack was expected to have a different degree of impact on the various travel markets. Here are some of the expected repercussions that were not borne out by the NHTS data:

• No significant decline in the percent of business air travel. This can be seen in table B1, appendix B, as the difference in pre-9/11 and post-9/11 business air travel was not statistically significant at the
• 0.05 level. There was no significant decline in the percent of business travel in general. The Pre- and Post-9/11 Econometric Analysis section discusses economic factors consistent with this finding.
• No significant decline in overall trips by older Americans. Older Americans are more risk averse and might be expected to travel proportionately less than other age groups when faced with the risk of terrorist attacks. However, those 65 and older did not show any statistically significant difference between pre-9/11 and post-9/11 when all modes were combined (table B2, appendix B).
• Little reduction in female long-distance
• travel patterns. It was hypothesized that females tend to be more risk averse than males and thus might show a decline in their proportion of long-distance travel after 9/11. However, female travel as a portion of all long-distance travel, with the exception of bus trips, showed no significant reduction (table B3, appendix B).
• No change in the percent of air trips from individuals living in urban areas, potentially the most affected by the 9/11 attacks, as compared to the percent of air trips from individuals living in rural areas. See table B7, appendix B, which shows urban travel by mode for pre-and post-9/11 travel. No significant change in percent of air trips from individuals living in urban areas was noted.

Conclusions

Without controlling for seasonality and economic effects, the NHTS data indicate that there was a decline in long-distance travel after 9/11. There was a decrease in the percentage of travelers traveling for international travel and personal business after 9/11. The decrease in travel for international and personal business occurred for both air and all modes combined. In addition, the proportion of trips under 100 miles increased after 9/11.

¹ National Bureau of Economic Research, Department of Commerce, The NBER Business-Cycle Dating Procedure, October 21, 2003, available at http://www.nber.org/cycles.

² Not all figures display the same modes. Only the modes that were found to have statistically significant differences between pre- and post- 9/11 data for a particular variable are shown.