OVERVIEW OF THE RESEARCH AND TASKS

The potential transportation impacts of telecommuting, defined here as salaried employees working from home instead of commuting to the usual workplace at the usual time, have been the subject of considerable research. A number of empirical studies on the transportation impacts of telecommuting are unanimous in finding substantial net reductions in vehicle miles traveled by telecommuters, on days that they telecommute, for as long as they are telecommuting (although it has also been suggested that the net impacts on a systemwide level will be rather modest). Because of its apparent transportation benefit, telecommuting has found its way into numerous public policy statements promoting trip reduction strategies – such as the United States bill H.R. 4475 (Sec. 359), signed into law on October 23, 2000, which states that "each executive agency [of the Federal government] shall establish a policy under which eligible employees of the agency may participate in telecommuting to the maximum extent possible without diminished employee performance".

However, all of the empirical studies to date have focused on short-term impacts, within one to two years of the adoption of telecommuting by the individual and/or the organization. In contrast, almost nothing is known about the long-term effects of telecommuting on activities such as residential location. A number of researchers have raised the issue that the ability to telecommute may prompt workers to move farther away from their jobs to cheaper or higher-amenity residential locations, resulting in a longer but less frequent commute. The total commute distance traveled in such cases may actually be higher after telecommuting than before. If such cases are quite common, then the degree to which telecommuting is promoted as a travel-reduction strategy should be re-examined. Hence, we need information on the extent and nature of the impact of telecommuting on residential location in order to inform transportation policies relating to telecommuting.

Several studies have explored the potential impacts of telecommuting on residential location and urban form from a theoretical economic perspective. Not surprisingly, all of these studies find that telecommuting can lead to further decentralization of the city. None of these theoretical approaches has been empirically validated, however, and empirical evidence with respect to impacts of telecommuting on residential location remains scanty. Of course, even if longer one-way commutes do result from telecommuting in some specific instances, the net change in commuting could still be negative if (1) at the individual level, the worker telecommutes often enough to compensate for the longer commute, and/or (2) at the aggregate level, the increases for some are outweighed by the reductions for others. Thus, it is important not to focus exclusively on specific adverse cases, but rather to examine the net or average impact across a pool of telecommuters, as we do in the current study.

One obvious reason for the lack of empirical studies of this relevant question is that residential location decisions take place over a longer time scale than the one- or two-year durations of most telecommuting pilot programs and their evaluations. A study would ideally need to follow telecommuters and otherwise comparable non-telecommuters over a period of perhaps 5-10 years in order to garner a large enough sample of moves from which to draw statistically robust conclusions, and few funding agencies (not to mention researchers) are willing to commit the necessary resources for that long a period of time.

To help at least partially bridge this gap in knowledge, the present study takes a more practical approach. Specifically, we conducted a single cross-sectional survey of current, former, and non-telecommuters, but asked retrospectively about their residential and job relocations, as well as their telecommuting engagement, over the preceding 10 years. Although such an approach is obviously subject to recall errors, we believe that residential and job changes are sufficiently rare and important events for most people, that their timing can be remembered with relative accuracy, particularly to within a given quarter of the year, which is the scale on which we captured the information. (Certainly, any number of demographic studies have been conducted using a similar method and based on the same assumption). Telecommuting engagement is probably less important to many people (and may change more often), and hence may be remembered less accurately, than residential or job changes. Nevertheless, respondents did not appear to have much trouble providing the requested information (there was relatively little apparently missing or unclear data), and we believe that despite inevitable imperfections in the data, analyzing the patterns they represent can offer useful insight into the complex relationships of interest. In any event, to our knowledge this study represents the first attempt to empirically study relationships among telecommuting, residential location, and total commute travel over a 10-year period.

A key challenge of analyzing specifically the impact of telecommuting on residential location is the proper attribution of causality. People move closer to or farther from work all the time, for reasons having nothing to do with telecommuting or even commuting. Given general trends toward urban decentralization in the US, and some evidence that commute lengths are increasing over time, a simple finding that telecommuters are moving farther away from work on average is not at all conclusive evidence that telecommuting should actually be “blamed” for such moves. Directly asking telecommuters whether that is the case is unreliable (although still worth doing), since respondents to a survey on telecommuting are likely to put undue emphasis on the role of telecommuting in their self-reported behavior. Comparing moves of telecommuters to those of an otherwise similar non-telecommuting group offers more rigorous evidence, but it can be difficult to identify a control group that is sufficiently similar to the telecommuters on key variables of interest.

In any case it should be kept in mind that telecommuters would not move farther from work simply because they are able to telecommute. Rather, in the worst case telecommuting simply makes possible a move that is desired for other reasons – to accommodate household needs, to obtain a larger and/or cheaper home or lot, to be near scenic locations or other amenities, and so on. Thus, telecommuting can be at worst viewed as a facilitator, not an actual driver, of such moves. Nevertheless, to the extent telecommuting is responsible for releasing a constraint that was preventing further decentralization (and to the extent such decentralization is considered socially undesirable), telecommuting is arguably as culpable as if it were a driver in its own right. In reality, of course, it could be rather difficult to disentangle the relative roles of drivers and facilitators, to determine the extent to which one factor could be singled out as being responsible for the move, as opposed to a number of factors acting in concert, with indivisible impact.

In view of these caveats with respect to attributing causality, the main focus of the present study is on the joint impact of telecommuting, residential location, and job location on transportation – specifically, commute travel. One immediate virtue of this focus is that it begs the question of causality entirely. We need not determine whether a relocation was an effect of telecommuting, a cause of it, or had any relationship to it at all. We simply compare the commute distance traveled of telecommuters and non-telecommuters. From the transportation planning perspective, this is arguably the key issue. Even if telecommuting does motivate some individuals to relocate far from work, if their commute frequency declines so much that their travel is still reduced, or if increases in their travel are outweighed by decreases in travel for other telecommuters, then policymakers may still be appropriately inclined to promote it.

KEY RESULTS

Key findings include: (a) One-way commute distances are higher for telecommuters than for non-telecommuters, consistent with prior empirical evidence and with expectation. (b) Average telecommuting frequency declines over time. This again is behaviorally plausible and consistent with prior independent results, but the trend seen here may be exaggerated due to the sampling bias in this study. (c) The first two findings notwithstanding, the average quarterly per-capita total commute distances are generally lower for telecommuters than for non-telecommuters, indicating that they telecommute often enough to more than compensate for their longer one-way commutes.

Weighting the sample to represent telecommuters proportionally to their presence in the national workforce, we find that on average, the proportion of telecommuters’ contribution to commute PMT is only 85% of their proportion in the workforce. In the last three quarters of the study, telecommuters are contributing to total commute PMT almost exactly in proportion to their representation in the workforce. Yet because of their longer one-way commutes, if telecommuters were not telecommuting but all else were held constant, their contribution to commute PMT would be 48% higher than their proportion in the workforce on average – 41% higher in the final quarter of the study period.

It should be emphasized again that in this study we do not argue for any particular direction of causality. That is, we are not able, on the basis of the analyses presented here, to discern whether longer commute distances encourage telecommuting, or conversely, whether the adoption of telecommuting facilitates residential relocations farther away from the workplace. As a corollary, we are unable to assess the desirability of telecommuting from this specific standpoint. While each direction of causality is possibly true for different people (or for the same people at different times), it is quite possible that the benign effects of telecommuting are dominant. The argument for this is that given a residential location, the adoption of telecommuting is virtually costless, while moving to a more distant location is a costly decision that usually involves a number of drives with a bigger weight than the possibility to telecommute. However, confirming this speculation would require an investigation into the role of telecommuting in relocation decisions that is beyond the scope of this paper.

Regardless of the direction of causality, however, the net outcome of decisions on telecommuting and residential/job location appears to be positive with respect to commute PMT. These results offer support for the continued promotion of telecommuting as a transportation demand management strategy, while also underlining the need for continued investigation into the complex relationships in which it plays a role.

Mokhtarian, Patricia L. and Gustavo O. Collantes (2003) Telecommuting, Residential Location, and Commute Distance Traveled: Evidence from State of California Employees. Paper presented at the NECTAR (Network on European Communication and Transportation Activities Research) conference: A New Millennium – Are Things the Same? UmeD, Sweden. June 13-15, 2003; and at the conference on City.Net: Cities in the Age of Telecommunications, Bauhaus University, Weimar, Germany. June 18-20, 2003.

Collantes, Gustavo O. and Patricia L. Mokhtarian (2003) Telecommuting and Residential Location: Relationships with Commute Distance Traveled for State of California Workers. Research Report No. UCD-ITS-RR-03-xx, Institute of Transportation Studies, University of California, Davis, in progress.

Gertz, Carsten and Patricia L. Mokhtarian (1999) Endbericht: Die Langfristigen Auswirkungen von Telearbeit auf das Verkehrsverhalten und Umzugsentscheidungen (The Long-Term Effects of Telecommuting on Travel Behavior and Residential Location). Report to BMW AG, Verkehrskonzepte Munchen, Germany, July (in German).