Telework-to-Play or Play-to-Telework? Investigating the directional relationship between telework and nonwork travel

• The factors influencing telework intensity and nonwork stops are examined. • Nonwork episodes are disaggregated into maintenance and leisure stops. • A joint cross-sectional multivariate model system is employed. • Latent segmentation is used to identify different causal directions. • Policy implications of the results are discussed. This study explores the intricate interrelationship between teleworking and weekly participation in two types of nonwork activities – maintenance episodes and leisure episodes. We employ a latent segmentation model to categorize the population into two segments: (1) those for whom nonwork activity participation influences telework frequency (“players”), and (2) those for whom telework frequency impacts nonwork activity participation (“workers”). Within each segment, to reduce any spurious association effects, we model teleworking and nonwork stop-making as a package by accommodating correlations in unobserved factors across the outcomes. The data for this analysis is drawn from a 2021–2022 weekly travel survey of Minnesotan workers in the Twin City region. The results reveal significant heterogeneity in the causal directionality of effect, with nonwork activity participation influencing telework intensity levels for about two-thirds of individuals (the “player” or NT segment), and telework levels influencing nonwork activity participation for the rest (the “worker” or TN segment). This result is in contrast to earlier studies that have predominantly assumed that all individuals belong to the TN or “worker” segment, which according to our analysis, is actually the case only for a minority of the population. Our analysis also indicates that, rather than asking the typical question of “does telework increase or decrease stop-making?”, the more pertinent question is “how does the number of nonwork stops relate to the intensity of teleworking?”. Our findings reveal an inverted U-shaped curve, with the highest number of nonstops occurring when individuals telework a few days a month or about one day per week. This, combined with the finding from the Asmussen et al. (2024a) study, indicates that low levels of teleworking (a day per week or less) actually increases both commute vehicle miles traveled (VMT) and nonwork VMT. Therefore, to achieve VMT reduction through teleworking, promoting moderate to high levels of telework appears important.