Does new information technology change commuting behavior?

Abstract

We estimate the long-run causal effect of information technology, i.e., Internet and powerful computers, as measured by the adoption of teleworking, on average commuting distance within professions in the Netherlands. We employ data for 2 years—1996 when information technology was hardly adopted and 2010 when information technology was widely used in a wide range of professions. Variation in information technology adoption over time and between professions allows us to infer the causal effect of interest using difference-in-differences techniques combined with propensity score matching. Our results show that the long-run causal effect of information technology on commuting distance is too small to be identified and likely to be absent. This suggests that, contrary to some assertions, the advent of information technology did not have a profound impact on the spatial structure of the labor market.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2

Notes

  1. 1.

    Several governments have official Internet pages that endorse teleworking, for example http://www.telework.gov.au in Australia and http://www.telework.gov in the USA.

  2. 2.

    In a similar fashion, an employee who teleworks part-time during the day, to avoid peak period congestion, experiences a lower generalized commuting cost as well.

  3. 3.

    As a referee noted, a priori it is difficult to tell whether workers with short or long commutes will take up teleworking. The former group might feel a strong disutility from commuting in general, while the latter group might want to avoid the long commuting time.

  4. 4.

    A similar argument has been given for households by Hu (2016), where the commuting distance of teleworkers automatically affects the commuting distance of non-teleworkers within the same household.

  5. 5.

    For example, workers with larger residences are more likely to live further away from the workplace and are more likely to prefer to work from home.

  6. 6.

    Zhu (2012, 2013) analyzed the effect of teleworking adoption on teleworkers and employs “Internet use at home” as an instrument for teleworking. However, when unobserved professional abilities of employees are correlated with the use of information technology, such as the Internet, then the instrument is not valid. So, this strategy implicitly assumes the absence of such a relation. Our identification strategy avoids such a restrictive assumption.

  7. 7.

    Note that we will not assume that the teleworking incidents rate was zero in 1996. Our estimation approach allows for the possibility that employees were working from home in 1996.

  8. 8.

    Nonetheless, telework correlates highly with other measures of information technology. A recent survey in 2015 among Dutch companies across 38 industrial sectors showed that teleworking correlates with 0.82 with using computers and Internet at work and with 0.85 with have access to mobile Internet.

  9. 9.

    Multiple factors affect technology adoption by an employee, including idiosyncratic distaste for teleworking, managerial practices (for example, Yahoo! forbade teleworking in 2013) or certain characteristics of the labor market (see Mokhtarian 1998).

  10. 10.

    As we explain later, our method differs from “difference-in-differences propensity score matching” methodology (see, e.g., Hijzen et al. 2013), which relies on panel data to observe the same individuals or firms over time. We observe the same professions over time but use propensity score matching to account for differences in labor force composition within professions over time.

  11. 11.

    The main advantage of using a binary measure of teleworkers share is that it drastically reduces the effect of measurement error in our teleworking variable. It seems reasonable to expect the effect of teleworking on commuting to be stronger the larger is the share of adopters within the profession. We confirm this assertion by repeating the entire analysis using data on professions with relatively low adoption rates, as we find no effect of technology on commuting.

  12. 12.

    In our application, we will use the years 1996 and 2010.

  13. 13.

    Treated and non-treated professions comprise of many different professions (both comprise professions which require low, medium or high education), so we allow for the possibility that some professions have a different trend, but not the average (industry-specific) trend.

  14. 14.

    We formally test the balancing property of the matching to insure that observations in two matched groups are similar in observed variables.

  15. 15.

    For an overview of the method and its limitations, we refer to Dehejia and Wahba (1999), Angrist and Pischke (2008) and Caliendo and Kopeinig (2008).

  16. 16.

    We will perform robustness checks with respect to the matching procedure.

  17. 17.

    In case the cohort effect is not important or that it does not differ across treated and non-treated professions, panel data would be preferable.

  18. 18.

    We consider employed individuals, between 18 and 64 years, working more than 12 h per week, with a one-way commute distance of less than 100 km.

  19. 19.

    When changes of distance within municipalities are in the same direction as changes across municipalities, which is plausible, our estimates of changes of commuting distances over time will be an underestimate. As long as treated and non-treated professions change similarly within municipalities, this would not affect our results. To check whether it does, we excluded those who work and live within the same municipality, but this does not change our main result qualitatively. But, as one referee pointed out, strictly speaking we look at changes in commuting distances between municipalities.

  20. 20.

    The self-reported average commuting distance in the Netherlands is 17 km over the period 2000–2008 (Groot et al. 2012). The difference with our data stems largely from within-municipality commutes that are assumed to be zero in our approach.

  21. 21.

    The original question in the Dutch language is “Waar werkt u in deze werkkring doorgaans?” which one might translate as “Where do you usually work on this job?” This question in the survey is clearly distinguished from another question about instances of overtime work at home.

  22. 22.

    See for definitions of teleworking as well Handy and Mokhtarian (1995), Mokhtarian et al. (1995), Pratt et al. (2000), Helling and Mokhtarian (2001), Mokhtarian et al. (2005) and Andreev et al. (2010).

  23. 23.

    For an overview on the measurement of teleworking, we refer to Sullivan (2003).

  24. 24.

    In choosing the scale of aggregation one has to trade off homogeneity of the resulting groups with the availability of the observations, to avoid classifying each employee as a single representative of a profession.

  25. 25.

    As a robustness check, we also provide results that are derived for a cutoff value of 10%.

  26. 26.

    Commuting within a municipality occurs in 49% (54%) and 31% (33%) in 2010 (1996) of, respectively, non-treated and treated professions.

  27. 27.

    For the propensity score matching estimates (and subsequent testing of the results), we use the “psmatch2” and “pstest” commands in Stata, developed by Leuven and Sianesi (2003).

  28. 28.

    The standardized bias is defined as the percentage difference of sample means in the treated and matched control subsamples as a percentage of the square root of the average of sample variances in both groups (Caliendo and Kopeinig 2008; Leuven and Sianesi 2003).

  29. 29.

    For example, we have included interaction and higher-order terms, excluded some variables and applied a more restrictive definition of treated professions. We have also applied other matching procedures such as one-to-one, three-to-one and five-to-one neighbors, and caliper, have varied the kernel bandwidth and performed estimations without replacements.

  30. 30.

    As one referee noted, this assumes that firms are equally likely to be sampled in our database. If firms that adopt ICT substitute labor for capital, then this might lead to a sample selection bias.

References

  1. Anas A, Arnott R, Small KA (1998) Urban spatial structure. J Econ Lit 36(3):1426–1464

    Google Scholar 

  2. Andreev P, Salomon I, Pliskin N (2010) State of teleactivities. Transp Res Part C Emerg Technol 18(1):3–20

    Article  Google Scholar 

  3. Angrist JD, Pischke JS (2008) Mostly harmless econometrics: an empiricist’s companion. Princeton University Press, Princeton

    Google Scholar 

  4. Arnold JM, Javorcik BS (2009) Gifted kids or pushy parents? foreign direct investment and plant productivity in Indonesia. J Int Econ 79(1):42–53

    Article  Google Scholar 

  5. Audirac I (2005) Information technology and urban form: challenges to smart growth. Int Reg Sci Rev 28(2):119–145

    Article  Google Scholar 

  6. Bailey DE, Kurland NB (2002) A review of telework research: findings, new directions, and lessons for the study of modern work. J Organ Behav 23(4):383–400

    Article  Google Scholar 

  7. Bloom N, Liang J, Roberts J, Ying ZJ (2015) Does working from home work? Evidence from a chinese experiment. Q J Econ 165:218

    Google Scholar 

  8. Cairncross F (1997) The death of distance: How the communications revolution will change our lives. Harvard Business School Press, Brighton

    Google Scholar 

  9. Caliendo M, Kopeinig S (2008) Some practical guidance for the implementation of propensity score matching. J Econ Surv 22(1):31–72

    Article  Google Scholar 

  10. Commander S, Harrison R, Menezes-Filho N (2011) Ict and productivity in developing countries: new firm-level evidence from Brazil and India. Rev Econ Stat 93(2):528–541

    Article  Google Scholar 

  11. De Borger B, Wuyts B (2011) The tax treatment of company cars, commuting and optimal congestion taxes. Transp Res Part B Methodol 45(10):1527–1544

    Article  Google Scholar 

  12. de Vos D, Meijers E, van Ham M (2018) Working from home and the willingness to accept a longer commute. Ann Reg Sci 61:375–398

    Article  Google Scholar 

  13. Dehejia RH, Wahba S (1999) Causal effects in nonexperimental studies: reevaluating the evaluation of training programs. J Am Stat Assoc 94(448):1053–1062

    Article  Google Scholar 

  14. Florida R (2005) The world is spiky globalization has changed the economic playing field, but hasn’t leveled it. Atl Mon 296(3):48

    Google Scholar 

  15. Friedman TL (2006) The world is flat [updated and expanded]: a brief history of the twenty-first century. Macmillan, London

    Google Scholar 

  16. Gaspar J, Glaeser EL (1998) Information technology and the future of cities. J Urban Econ 43(1):136–156

    Article  Google Scholar 

  17. Girma S, Görg H (2007) Evaluating the foreign ownership wage premium using a difference-in-differences matching approach. J Int Econ 72(1):97–112

    Article  Google Scholar 

  18. Glaeser EL (2008) Cities, agglomeration, and spatial equilibrium. OUP Catalogue, Oxford

    Google Scholar 

  19. Groot SPT, de Groot HLF, Veneri P (2012) The educational bias in commuting patterns: micro-evidence for the Netherlands. Tinbergen institute discussion paper 12-080/3. https://doi.org/10.2139/ssrn.2119929

  20. Handy SL, Mokhtarian PL (1995) Planning for telecommuting measurement and policy issues. J Am Plan Assoc 61(1):99–111

    Article  Google Scholar 

  21. Helling A, Mokhtarian PL (2001) Worker telecommunication and mobility in transition: consequences for planning. J Plan Lit 15(4):511–525

    Article  Google Scholar 

  22. Hijzen A, Martins PS, Schank T, Upward R (2013) Foreign-owned firms around the world: a comparative analysis of wages and employment at the micro-level. Eur Econ Rev 60:170–188

    Article  Google Scholar 

  23. Hu L (2016) Association between telecommuting and household travel in the Chicago metropolitan area. J Urban Plan Dev 142(3):04016,005

    Article  Google Scholar 

  24. IDS (1996) IDS Study 616. Income Data Services

  25. James A (2014) Work-life ‘balance’ and gendered (im) mobilities of knowledge and learning in high-tech regional economies. J Econ Geogr 14(3):483–510

    Article  Google Scholar 

  26. Jorgenson DW, Ho M, Stiroh K (2008) A retrospective look at the us productivity growth resurgence. J Econ Perspect 22(1):3–24

    Article  Google Scholar 

  27. Keynes JM (2010) Economic possibilities for our grandchildren. In: Essays in persuasion. Palgrave Macmillan, London, pp 321–332

  28. Kim SN (2016) Two traditional questions on the relationships between telecommuting, job and residential location, and household travel: revisited using a path analysis. Ann Reg Sci 56(2):537–563

    Article  Google Scholar 

  29. Kim SN, Mokhtarian PL, Ahn KH (2012) The Seoul of Alonso: new perspectives on telecommuting and residential location from South Korea. Urban Geogr 33(8):1163–1191

    Article  Google Scholar 

  30. Leuven E, Sianesi B (2003) PSMATCH2: Stata module to perform full Mahalanobis and propensity score matching, common support graphing, and covariate imbalance testing. Statistical software components S432001, Boston college department of economics. Revised 1 Feb 2018

  31. Lund JR, Mokhtarian PL (1994) Telecommuting and residential location: theory and implications for commute travel in monocentric metropolis. Transp Res Rec 1463:10–14

    Google Scholar 

  32. McCann P (2008) Globalization and economic geography: the world is curved, not flat. Camb J Reg Econ Soc 1(3):351–370

    Article  Google Scholar 

  33. Mokhtarian PL (1998) A synthetic approach to estimating the impacts of telecommuting on travel. Urban Stud 35(2):215–241

    Article  Google Scholar 

  34. Mokhtarian PL, Handy SL, Salomon I (1995) Methodological issues in the estimation of the travel, energy, and air quality impacts of telecommuting. Transp Res Part A Policy Pract 29(4):283–302

    Article  Google Scholar 

  35. Mokhtarian PL, Collantes GO, Gertz C (2004) Telecommuting, residential location, and commute-distance traveled: evidence from state of california employees. Environ Plan A 36(10):1877–1897

    Article  Google Scholar 

  36. Mokhtarian PL, Salomon I, Choo S (2005) Measuring the measurable: Why cant we agree on the number of telecommuters in the us? Qual Quant 39(4):423–452

    Article  Google Scholar 

  37. Moos M, Skaburskis A (2008) The probability of single-family dwelling occupancy comparing home workers and commuters in canadian cities. J Plan Educ Res 27(3):319–340

    Article  Google Scholar 

  38. Mulalic I, Van Ommeren JN, Pilegaard N (2014) Wages and commuting: quasi-natural experiments’ evidence from firms that relocate. Econ J 124(579):1086–1105

    Article  Google Scholar 

  39. Paoli P (2001) Third European survey on working conditions 2000. Office for official publications of the European Communities

  40. Pissarides CA (2000) Equilibrium unemployment theory. MIT Press, Cambridge

    Google Scholar 

  41. Pratt JH, et al (2000) Telework and society-implication for corporate and societal cultures. In: Century. Xavier University, Citeseer

  42. Rhee HJ (2008) Home-based telecommuting and commuting behavior. J Urban Econ 63(1):198–216

    Article  Google Scholar 

  43. Rosenbaum PR, Rubin DB (1985) Constructing a control group using multivariate matched sampling methods that incorporate the propensity score. Am Stat 39(1):33–38

    Google Scholar 

  44. Safirova E (2002) Telecommuting, traffic congestion, and agglomeration: a general equilibrium model. J Urban Econ 52(1):26–52

    Article  Google Scholar 

  45. Sang S, OKelly M, Kwan MP (2011) Examining commuting patterns: results from a journey-to-work model disaggregated by gender and occupation. Urban Stud 48(5):891–909

    Article  Google Scholar 

  46. Stiebale J, Trax M (2011) The effects of cross-border M&As on the acquirers’ domestic performance: firm-level evidence. Can J Econ 44(3):957–990

    Article  Google Scholar 

  47. Storper M, Venables AJ (2004) Buzz: face-to-face contact and the urban economy. J Econ Geogr 4(4):351–370

    Article  Google Scholar 

  48. Sullivan C (2003) What’s in a name? Definitions and conceptualisations of teleworking and homeworking. New Technol Work Employ 18(3):158–165

    Article  Google Scholar 

  49. Van Ommeren J, Rietveld P, Nijkamp P (1999) Job moving, residential moving, and commuting: a search perspective. J Urban Econ 46(2):230–253

    Article  Google Scholar 

  50. Welz C, Wolf F (2010) Telework in the European union. European Foundation for the Improvement of Living and Working Conditions, Dublin

  51. Zax JS (1991) The substitution between moves and quits. Econ J 101(409):1510–1521

    Article  Google Scholar 

  52. Zhu P (2012) Are telecommuting and personal travel complements or substitutes? Ann Reg Sci 48(2):619–639

    Article  Google Scholar 

  53. Zhu P (2013) Telecommuting, household commute and location choice. Urban Stud 50(12):2441–2459

    Article  Google Scholar 

Download references

Acknowledgements

Financial support from The Netherlands Organization for Scientific Research (NWO) is gratefully acknowledged. This paper is part of TRISTAM Project (Traveler Response and Information Service Technology—Analysis and Modeling).

Author information

Affiliations

Authors

Corresponding author

Correspondence to Thomas de Graaff.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Appendix

Appendix

See Fig. 3 and Table 6.

Fig. 3
figure3

Empirical distribution of shares of teleworkers within professions in 2010. (Note: The share of professions with no teleworkers is 0.53 and not shown here)

Table 6 Descriptive statistics

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Gubins, S., van Ommeren, J. & de Graaff, T. Does new information technology change commuting behavior?. Ann Reg Sci 62, 187–210 (2019). https://doi.org/10.1007/s00168-018-0893-2

Download citation

JEL Classification

  • J22
  • R23
  • R41