Walls, Nelson, Safirova
Telecommuting and Environmental Policy
stakeholders outlined several important goals and objectives to be achieved in the program, including
the following:
• ascertaining the viability of developing, testing and promoting economic incentives, specifically
including emissions credits that may be tradable;
• building partnerships within U.S. communities to address air quality, congestion and quality of
life issues associated with teleworking;
• defining and sharpening regulatory and statutory issues necessary to promote air emissions
trading incentives, including mechanisms for cross-sector emissions trading;
• developing and evaluating methods for calculating reductions in emissions of precursors of
ground level ozone and greenhouse gases achieved as the result of vehicle miles traveled; and
• gathering important data and information about transportation, urban planning, and the
environmental impacts of telework.
The original five metropolitan areas in the program were Chicago, the District of Columbia,
Houston, Los Angeles, and Philadelphia. Chicago dropped out of the program in 2000 and was replaced
by Denver. The “ecommute” program commenced in June 2001 and by March 2004, 49 companies
with 535 employees had participated.
The collection of data on teleworking activity in the ecommute program, as well as information
on other commuting modes, is done using Teletrips, a web-based commute-tracking software developed
by the Canadian-based firm, Teletrips Inc. The software was designed specifically to track the emission
benefits from telecommuting and other “green” commute alternatives (such as carpooling and biking).
Participating employees fill out a form at the end of each week indicating for each day whether or not
they went to work and their means of travel. The interface is designed to be user-friendly and
participants report that the weekly log can be filled out in under a minute. Reporting is prompted by
weekly e-mail reminders.
Upon entering the program, each employee provides information about their vehicle and their
commute. Information about the employee’s vehicle includes the year, make, and model of the vehicle,
the engine type, and whether or not it has passed an emissions test. Commute information includes the
distance of the journey to work, as well as the time that the employee leaves home, arrives at work, and
leaves work, arrives at home. The employee reports his mode choice for each day, including driving
alone to work, carpooling, using public transit, walking, biking, telecommuting at home, or
telecommuting at a telework center. The employee also reports if he does not work on that particular
day.
Teletrips is not a full-fledged trip diary, and therefore it does not capture the total travel
behavior of individuals who telecommute. For example, it is impossible to determine from the data if
individuals drive more for non-work purposes on their telecommuting days, thus offsetting some of the
benefits associated with not driving to work. Furthermore, detailed information about the employees
enrolled in Teletrips is not available. So, for example, we do not know the income, marital status,
education, and so forth of these workers. We also do not have reliable information on the types of jobs
in which these workers are employed, their length of employment, or other job status information.
Another factor limiting the scope of conclusions that can be drawn from the Teletrips data is the
self-selected nature of the sample. Self-selection occurs on two levels. First, the decision to participate
in the program is made at the firm level, so the sample is limited to those firms that have an interest in
telecommuting and who happened to be contacted by the local organizations responsible for marketing
the program. Second, even within those firms, only employees who are interested in telecommuting
sign up. Therefore, information is collected about only those employees who are most likely to
participate in a telecommuting-tracking program like ecommute. Unfortunately, this data set does not
provide a suitable “control” group that would allow more general questions about telecommuting
behavior to be answered.
11