Abstract
This article investigates the selection of a distance measure in location modeling. While in the empirical literature the choice usually boils down to picking one single measure, this research proposes a flexible approach in which several measures may be used in parallel to capture the surrounding economic landscape. This is intended to acknowledge that interactions between agents may take several forms, occurring through different channels and as such being based on different measures. The methodology is applied to the location choice of establishments in the Paris region, using a mixture of ”mono-distance” hurdle-Poisson models. Seven distance measures are considered: Euclidean distance, the travel times by car (for the peak and off-peak periods) and by public transit, and the corresponding network distances. For all the economic sectors considered, the mixture of hurdle-Poisson models performs significantly better than the “pure” mono-distance models. This corroborates that local spatial spillovers are indeed channeled by different means, hence best represented using several measures. The combination of peak and off-peak road travel times (slightly) outperforms other combinations including the Euclidean distance, supporting the choice of meaningful over more abstract measures in spatial econometric models. The distance measure most likely to capture local spatial spillovers varies depending on the economic sector examined, reflecting differences between sectors in operations and location choice criteria.
Similar content being viewed by others
Notes
As noticed by Miller and Wentz (2003), many researchers adopt the Euclidean distance without realizing its underlying assumptions or its alternatives. This representation has been originally proposed due to scarce data and low computational power at the time, rather than because of its alleged universality. Yet, it remains largely applied to this day.
In the case of continuous data, spatial cross-regressive models have already been used to combine several measures (e.g., Kang and Dall’erba (2016)).
Nguyen et al. (2012) develops a relocation choice model using an average travel distance as a proxy for the distance among zones and firms.
In this paper, local spatial spillovers refer to the direct effect of the characteristics of other zones onto the profitability of a given zone. Following the terminology used in Márquez et al. (2010) or LeSage (2014), formally they are represented using spatially lagged explanatory variables as visible in Eq. 2. Endogenous spatial lags (such as in spatial auto-regressive models) are not considered.
One might set this parameter at a value smaller or larger than 1 to obtain stronger or weaker local spillover effects. Similarly, one may also allow μm,k to vary with the variables k and/or the measure m. These possibilities are left aside for further research.
More specifically, the basic specification (i.e. before the mixture) used in this paper might be regarded as a nonlinear counterpart of the spatial lag of X model (SLX) in the discrete case, with two specificities relatively to the standard SLX formulation (see, e.g., LeSage (2014)): the vector of spatially lagged explanatory variables WX is transformed with a logarithm, and a hurdle model is included.
The Paris region, also called Île-de-France, is a vibrant and innovative region with over 5.6 million jobs, 37 percent of national executives, and 40 percent of the national workforce in research and development. It is the 1st R&D hub in Europe and the third worldwide. 11.7 million people representing over 19 percent of the country population live in an area which covers only 2.2 percent of national grounds. The GDP of the region amounts to 31 percent of total French GDP (http://www.grand-paris.jll.fr/fr/paris/chiffres-cles/) that is 612 billion euros (2012). The Paris region is the 1st European region considering the number of firms classified in Fortune 500 (July 2014). The Paris region consists of 1300 municipalities encompassing the inner city of Paris and its suburbs. Yet, very large differences in population and employment densities are to be found between the Paris city and its outer periphery.
Distance matrices based on a transport network being relatively stable over time at a regional scale, especially in the Paris region where the transport networks are already well developed, this one year difference should have a very limited impact on our results.
These three sectors were chosen as best representative of the usefulness of the methodology as well as of the diversity of situations regarding which distance measure best captures spatial effects. For the other sectors, similar results were obtained, though the model improvement (in terms of BIC) was slightly less important. For a very few sectors, the mixture was also found not bring significant improvement. Detailed results for all sectors are available upon request from the authors.
This issue is closely related to the type of trip undertaken: short distances typically correspond to intra-urban trips, and longer distances to inter-urban trips.
While the detour factor well illustrates the proximity difference between two points when using either the Euclidean distance or a given network distance, it does not well reflect the level of spatial correspondence between the two corresponding paths, so that more suitable indicators should be used instead for these ends (see Matisziw and Demir (2016), on this issue). Here, the analysis focuses on the notion of proximity and does not intend to exploit differences at the path-level, so that the mean detour factor is informative enough.
All three combinations are based on sets of two measures (i.e., \(\vert \mathcal {M}\vert = 2\)). To keep the exposition of the results simple, larger sets (\(\vert \mathcal {M}\vert >2\)) or sets containing mixtures of the described measures were not considered. For instance, including linear combinations of the three measures considered would allow considering establishments that are not purely oriented toward peak/off-peak travel time or toward the crow-fly distance when approximating the profit level of a zone. While this represents a limitation of the present study, search for the optimal number of clusters and their exact definitions will be the object of future works.
Readers interested in this issue may refer to Buczkowska and de Lapparent (2017).
References
Acs ZJ, Audretsch D, Feldman MP (1994) R&D spillovers and recipient firm size. Rev Econ Stat 100(1):336–367
Anselin L (2003) Spatial externalities, spatial multipliers, and spatial econometrics. Int Reg Sci Rev 26(2):153–166
Arauzo-Carod JM, Liviano-Solís D, Manjón-Antolín M (2010) Empirical studies in industrial location choice: An assessment of their methods and results. J Reg Sci 50(3):685–711
Aten B (1997) Does Space Matter? International Comparisons of the Prices of Tradables and Nontradables. Int Reg Sci Rev 20:35–52
Axhausen KW (2003) Definitions and measurement problems. In: Axhausen KW, Madre J-L, Polak JW, Toint P (eds) Capturing long distance travel. Research Science Press, Baldock
Benezech V, Coulombel N (2013) The value of service reliability. Transp Res B Methodol 58:1–15
Bhat ChR, Paleti R, Singh P (2014) A spatial multivariate count model for firm location decisions. J Reg Sci 54(3):462–502
Billé AG, Benedetti R, Postiglione P (2017) A two-step approach to account for unobserved spatial heterogeneity Spatial Economic Analysis, in press
Bodson P, Peeters D (1975) Estimations of the coefficients in a linear regression in the presence of spatial autocorrelation: an application to a Belgian labour-demand function. Environ Plan A 7:455–72
Boscoe FP, Henry KA, Zdeb MS (2012) A nationwide comparison of driving distance versus straight-line distance to hospitals. Prof Geogr 64(2):188–196
Buczkowska S, de Lapparent M (2014) Location choices of newly-created establishments: Spatial patterns at the aggregate levels. Reg Sci Urban Econ 48:68–81
Buczkowska S, de Lapparent M (2017) Location choices under strategic interactions: Interdependence of establishment types. Working paper
Buczkowska S, Coulombel N, de Lapparent M (2018) The Paris region travel times and distances dataset. https://doi.org/10.7910/DVN/E85DBD, Harvard Dataverse, V1
Chalasani VS, Denstadli JM, Engebretsen Ø, Axhausen KW (2005) Precision of Geocoded locations and network distance estimates. J Transp Stat 8(2):1–15
Chatman DG, Noland RB, Klein NJ (2016) Firm births, access to transit, and agglomeration in Portland, Oregon, and Dallas, Texas. Transp Res Rec: J Transp Res Board 2598:1–10
Chen Z, Haynes K (2014) Spatial impact of transportation infrastructure: A spatial econometric CGE approach. In: Rose A, Nijkamp P (eds) Regional science matters? studies dedicated to walter Isard. Springer, Berlin
Cole JP, King CAM (1968) Quantitative geography. Wiley, London
Combes PP, Lafourcade M (2003) Core-periphery patterns of generalized transport costs: France, 1978-98 C.E.P.R. Discussion papers 3958
Combes PP, Lafourcade M (2005) Transport costs: measures, determinants, and regional policy implications for France . J Econ Geogr 5(3):319–349. Oxford University Press
Condeço-Melhorado A, Reggiani A, Gutiérrez J (2018) New data and methods in accessibility analysis. Netw Spatial Econ 18(2):237–240
Conley TG, Ligon E (2002) Economic distance and cross-country spillovers. J Econ Growth 7(2):157–187
Conn AR, Gould NIM, Toint PL (2000) Trust region methods. MPS-SIAM series on optimization, society for industrial and applied mathematics
Corrado L, Fingleton B (2012) Where is the economics in spatial econometrics? J Reg Sci 52(2):210–239
Coulombel N, Leurent F (2013) Les ménages arbitrent-ils entre coût du logement et coût du transport: une réponse dans le cas francilien. Economie & Statistique 457-458:57–75
de Dios Ortuzar J, Willumsen LG (2011) Modeling transport, 4th edn. Wiley
Dempster AP, Laird NM, Rubin DB (1977) Maximum likelihood from incomplete data via the EM algorithm. J R Stat Soc Ser B 39(1):1–38
Denain JC, Langlois P (1998) Cartographie en anamorphose. Mappemonde 49(1):16–19
DRIEA Île-de-France (2008) MODUS v21 Documentation détaillée du modéle de déplacements de la DREIF Technical report
Dubé J, Brunelle C, Legros D (2016) The review of regional studies location theories and business location decision: A micro-spatial investigation of a nonmetropolitan area in Canada. Rev Reg Stud 46:143–170
Duran-Fernandez R, Santos G (2014) Gravity, distance, and traffic flows in Mexico. Res Transp Econ 46:30–35
Duranton G, Overman HG (2005) Testing for localization using micro-geographic data. Rev Econ Stud 72(4):1077–1106
Eaton BC, Lipsey RG (1980) The block metric and the law of markets. J Urban Econ 7:337–347
Eliste P, Federiksson PG (2004) Does trade liberalization cause a race to the bottom in environment policies? A spatial econometric analysis. In: Anselin L, Florax R (eds) Advances in spatial econometrics. Springer, New York
Faber B (2014) Trade integration, market size, and industrialization: Evidence from China’s national trunk highway system. Rev Econ Stud 81(3):1046–1070
Fetter FA (1924) The economic law of market areas. Q J Econ 39:520–529
Fingleton B (2008) A generalized method of moments estimator for a spatial model with moving average errors, with application to real estate prices. Empir Econ 34:35–57
Fingleton B, Le Gallo J (2008) Estimating spatial models with endogenous variables, a spatial lag and spatially dependent disturbances: finite sample properties. Pap Reg Sci 87(3):319–339
Fransen K, Neutens T, De Mayer P, Deruyter G (2015) A commuter-based two-step floating catchment area method for measuring spatial accessibility of daycare centers. Health Place 32:65–73
Graham DJ (2007) Variable returns to urbanization and the effect of road traffic congestion. J Urban Econ 62(1):103–120
Gutiérrez J (2001) Location, economic potential and daily accessibility: an analysis of the accessibility impact of the high-speed line Madrid-Barcelona-French border. J Transp Geogr 9(4):229–242
Guy CM (1983) The assessment of access to local shopping opportunities: A comparison of accessibility measures. Environ Plann B 10:219–238
Haggett P (1967) Network models in geography. In: Chorley RJ, Haggett P (eds) Integrated models in geography. Methuen, London, pp 609–668
Kang D, Dall’erba S (2016) An examination of the role of local and distant knowledge spillovers on the US regional knowledge creation. Int Reg Sci Rev 39(4):355–385
Klier T, McMillen D (2008) Clustering of auto supplier plants in the United States. Bus Econ Stat Am Stat Assoc 26(4):460–471
Kwon Y (2002) Rent-commuting cost function versus rent-distance function. J Reg Sci 42(2):773–791
Lambert DM, Brown JP, Florax RJGM (2010) A two-step estimator for a spatial lag model of counts: theory, small sample performance and an application. Reg Sci Urban Econ 40(4):241–252
de Lapparent M, Koning M (2015) Travel discomfort-time tradeoffs in Paris subway: an empirical analysis using interval regression models EPFL-REPORT-204605
Le Gallo J, Dall’erba S (2008) Spatial and sectoral productivity convergence between European regions, 86, 1975-2000. Pap Reg Sci 87(4):505–525
LeSage JP (2014) What regional scientists need to know about spatial econometrics. Rev Reg Stud 44(1):13–32
LeSage JP, Pace RK (2008) Spatial econometric modeling of origin-destination flows. J Reg Sci 48(5):941–967
Liesenfeld R, Richard JF, Vogler J (2015) Likelihood evaluation of high-dimensional spatial latent gaussian models with non-gaussian response variables. Available at SSRN: SSRN-id2196041 2
Liviano-Solís D, Arauzo-Carod JM (2013) Industrial location and interpretation of zero counts. Ann Reg Sci 50:515–534
Márquez MA, Ramajo J, Ewings G (2010) Measuring the spillover effects of public capital: a bi-regional structural vector autoregressive analysis. Lett Spat Resour Sci 3:111–125
Matisziw TC, Demir E (2016) Measuring spatial correspondence among network paths. Geogr Anal 48:3–17
Miller HJ (2004) Tobler’s first law and spatial analysis. Ann Assoc Am Geogr 94(2):284–289
Miller HJ, Wentz EA (2003) Representation and spatial analysis in geographic information systems. Ann Assoc Am Geogr 93(3):574–594
Nguyen CY, Sano K, Tran TV, Doan TT (2012) Firm relocation patterns incorporating spatial interactions. Ann Reg Sci 50(3):685–703
Parent O, LeSage JP (2008) Using the variance structure of the conditional autoregressive spatial specification to model knowledge spillovers. J Appl Econ 23(2):235–256. John Wiley & Sons Ltd
Partridge MD, Rickman DS, Ali K, Olfert MR (2008) Lost in space: population growth in the American hinterlands and small cities. J Econ Geogr 8(6):727–757
Perreur J, Thisse J (1974) Central metrics and optimal location. J Reg Sci 14:411–421
Rietveld P, Zwart B, van Wee B, van den Hoorn T (1999) On the relationship between travel time and travel distance of commuters. Ann Reg Sci, Springer 33(3):269–287
Rincke J (2010) A commuting-based refinement of the contiguity matrix for spatial models, and an application to local police expenditures. Reg Sci Urban Econ 40:324–330
Rushton G, Thill JC (1989) The effect of distance metric on the degree of spatial competition between firms. Environ Plan A 21(4):499–507
Slade ME (2005) The role of economic space in decision making, ADRES Lecture. Annales D’Economie et de Statistique 77:1–20
Talen E, Anselin L (1998) Assessing spatial equity: an evaluation of measures of accessibility to public playgrounds. Environ Plan A 30:595–613
Tobler W (1970) A computer movie stimulating urban growth in the Detroit region. Econ Geogr 46:234–240
Vega SH, Elhorst JP (2013) On spatial econometrics models, spillovers effects, and W ERSA conference papers ersa13p222, European Regional Science Association
Veneri P (2010) Urban polycentricity and the costs of commuting: evidence from Italian metropolitan areas. Growth and Change 41(3):403–429
Weisbrod G (2008) Models to predict the economic development impact of transportation projects: historical experience and new applications. Ann Reg Sci 42(3):519–543
Acknowledgements
The authors would like to thank Professor Josep-Maria Arauzo-Carod and Professor James LeSage, all the anonymous reviewers, and the editor for their constructive suggestions and very helpful comments on the paper. We also thank the DRIEA-IF and Caliper for providing access to the MODUS model and the TransCAD software, respectively.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
An earlier version of this paper was presented at the 55th European Regional Science Association Congress (Lisbon, Portugal, August 25-29, 2015), 14th International Workshop on Spatial Econometrics and Statistics (Paris, France, May 27-28, 2015), and published as a Ecole Polytechnique Fédérale de Lausanne Technical Report. The usual disclaimers apply.
Appendix
Appendix
Rights and permissions
About this article
Cite this article
Buczkowska, S., Coulombel, N. & de Lapparent, M. A comparison of Euclidean Distance, Travel Times, and Network Distances in Location Choice Mixture Models. Netw Spat Econ 19, 1215–1248 (2019). https://doi.org/10.1007/s11067-018-9439-5
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11067-018-9439-5