Communication technology and travel demand models

Abstract: Transportation planners have traditionally focused onphysical travel only, and disregarded the fact that other modesof communication may influence travel demand. However, moderntelecommunications are rapidly increasing the accessibility toactivities that previously only could be reached by physicaltransportation. This development calls for methods to analyseinteractions between telecommunications and transport systems.The objective of this thesis is to accomplish a betterunderstanding of if and how impacts of information technologycould be implemented in travel demand models. An important partof this issue is to investigate what kind of data that isneeded.This thesis also aims at investigating whether theCommunication Survey, KOM, collected by Swedish Institute forTransport and Communications, SIKA, can be used to improvetransport modelling with respect to how moderntelecommunications influence travel demand. KOM is a one-daytravel and communication diary survey, including information onthe respondents telecommuting habits as well as socio-economicstatus. One problem was the small sample size in KOM, whichmade the analyses uncertain. Since KOM is collected on a yearlybasis, it is still possible to apply similar analysis methodswithin a few years, using a larger data set, which might enableextended analyses. The small sample in KOM available to date isbest suited for general descriptive analyses of communicationpatterns in Sweden. The main conclusions of the paper aretherefore connected to the methods and future datacollection.The thesis includes three papers. The first paper tested amodel approach that assumes substitution between travel andnon-travel based communication, using the KOM database. Traveldemand models are in general constructed as nested logit modelswith frequency, mode and destination choice levels. In thepaper, non-travel based modes of communication were included inthe choice set of such a model. The non-travel based modes ofcommunication considered were Internet (and e-mail), ordinarymail and telephone contacts. The model was developed for postand bank activities only, since that was the only activity forwhich the numbers of contacts and trips were large enough toallow model estimation. Several conclusions could be drawn.Describing the utility of the non-travel based alternatives isdifficult and needs more research. The analysisis also verysensitive to how activities are defined. It is furtheressential that the data collection is more process orientedthan traditional cross-sectional data is when analysing traveland telecommunications interactions. That is, habits ofperforming particular activities, including both trips anddifferent types of contacts, must be studied. The second andthird papers investigate telecommuting. As a first step toreach the goal of forecasting telecommuting, the second paperexamined the characteristics of current telecommuters by use ofKOM. This was mainly accomplished by estimating a telecommutingadoption model of logit type. However, only 122 employees outof 7578 actually telecommutes full days at home. Thesetelecommuters work primarily in information- and service-basedindustrial sectors concerned with computers, finance orcommunication. The difficulties in describing the utility ofthe telecommunications based alternatives (representing?no travel?) concerned also the telecommutingadoption model. Also impacts on travel from telecommuting wereinvestigated. Comparing the average commuting distance showedthat employees who exclusively telecommute full days havelonger commuting distances than others, but that othertelecommuters do not have longer average commuting distances.Telecommuting in general does not seem to be influenced by lowaccessibility to the labour market.The third paper used data collected from a working sitewithin the company Ericsson, located in the office district ofNacka Strand in Stockholm during the autumn 2002. Thetelecommuting frequency was substantially higher at Ericssonthan in the workforce as a whole. The propensity to adopttelecommuting was modelled as a function of socio-economicvariables and access to technical equipment, work tasksuitability and management attitudes, as perceived by theemployees. The focuswas to identify tools that the company canuse to promote telecommuting, and to find incentives for thecompany to promote telecommuting. Technical equipment, suitablework tasks and managers attitude were identified as constraintsfor telecommuting. The employees also perceived that theybecame more efficient and saved time when telecommuting.