Whether telecommuting and personal travel are complements or substitutes is a key question in urban policy analysis. Using survey responses on telecommuting, previous empirical studies relied on small regional samples and concluded that telecommuting is more likely to function as a substitute for commute travel. These studies also agreed that the substitution effect was small.Using data from the 2001 and 2009 National Household Travel Surveys (NHTS), this study involves two large national samples to try to identify the impact of telecommuting on workers' travel patterns. Through a series of empirical tests, this research investigates how telecommuting influences workers' one-way commute trips, daily total work trips and daily non-work trips, and tries to provide some answers to a question that has been discussed for some years-namely, whether telecommuting and personal travel are complements or substitutes. The results of these tests suggest that telecommuting has been an important factor in shaping personal travel patterns over the 2001-2009 period, and that telecommuting indeed has a complementary effect on not just workers' one-way commute trips but also their daily total work trips and total non-work trips.

To understand whether telecommuting could be part of the policy solutions for greenhouse gas reduction in the transportation sector, this study uses Instrumental Variable Tobit models and data from 2001 and 2009 National Household Travel Surveys to explore whether telecommuting reduces or increases the daily work and non-work vehicle miles traveled. Our findings suggest telecommuters have more vehicle miles traveled for both daily work and non-work trips than non-telecommuters. Adding the findings that telecommuting has no impact on other non-working household member's daily total (nonwork) trips, we can possibly argue that households with telecommuter(s) tend to have higher daily total vehicle miles traveled. Our estimated marginal effect of telecommuting on worker's daily total trips indicates that, a telecommuter on average travels 38 vehicle miles more on a daily basis in 2001 and 45 vehicle miles more in 2009 compared with a non-telecommuter. These increases in vehicle miles traveled translate into a rather large increase in greenhouse gas emissions in the U.S. equivalent to adding 7,248,845 cars in 2001 and 8,808,165 in 2009 to the road. Moreover, the difference of this marginal effect between 2001 and 2009 suggests the impact of telecommuting on worker's daily total vehicle miles traveled had increased over time. With the emerging work arrangements to work from home, telecommuting has been welcomed in this changing environment, not only by individual workers and employers but also policymakers. But the outcomes seem to be opposite to what policy makers may have expected for GHG emission reductions.

Previous empirical studies have made contributions to the understanding of the impact of telecommuting on individual travel patterns. There has been much less research that has examined the impact of telecommuting on commute travel at the household level. Using data from the 2001 and 2009 US National Household Travel Surveys, this study focuses on one-worker and two-worker households and investigates how telecommuting affects household one-way commute distance and duration. The results show that telecommuting increases the commute distance and duration for both one-worker households and two-worker households. It is also found that, in two-worker households, the telecommuting status of one worker does not increase the commute distance and duration of the other worker. These findings suggest that telecommuting (two-worker) households tend to choose locations involving a longer total one-way commute than non-telecommuting households, and this difference is largely due to the longer commute of their telecommuting members