1 / 15

Valerio Gatta * and Edoardo Marcucci * * DIPES/CREI, University of Roma Tre

Urban freight distribution policies: joint accounting of non-linear attribute effects and discrete mixture heterogeneity. Valerio Gatta * and Edoardo Marcucci * * DIPES/CREI, University of Roma Tre.

karim
Download Presentation

Valerio Gatta * and Edoardo Marcucci * * DIPES/CREI, University of Roma Tre

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Urban freight distribution policies: joint accounting of non-linear attribute effects and discrete mixture heterogeneity Valerio Gatta* and EdoardoMarcucci* * DIPES/CREI, University of Roma Tre "Transport, Spatial Organization and Sustainable Economic Development” - Venice - September 18-20, 2013 XV Conference of the Italian Association of Transport Economics and Logistics

  2. Outline • Research goals • Survey and Data description • Main results • Conclusions

  3. Research goals • Study context • Urban Freight Transport (UFT). Main agents: retailers, transport providers, own account • Policy makers are interested in knowing, before implementing a given policy, the most likely reactions • One-size-fit-all policies are usually implemented with mixed results • Contributions to UFT literature • Lack of appropriate data (elicitation costs & low interest of agents ► in-depth investigation of transport providers’ preferences • Policy makers usually evaluate policies assuming linear effects on agent’s utility for attribute variations. • Not only inter-agent but also intra-agent heterogeneity • Joint analysis of heterogeneity & non-linear effects

  4. Survey and data description (I) • Project • Stated Ranking Exercise in Rome’s Limited Traffic Zone • Volvo Research Foundation (2009), “Innovative solutions to freight distribution in the complex large urban area of Rome” • Main steps • Advancement from stakeholder consultation to final attribute selection criteria • Attribute definition • Levels and ranges selection • Progressive design differentiation by agent-type with updated priors (efficient design, 3+1 waves)

  5. Survey and data description (II) • Attribute levels and ranges • Example of a ranking task

  6. Survey and data description (III) • The sample of transport providers • 66 units • Total number of observations: 1128 Transport provider agent distribution by main freight sector • Food(fresh, hotels, restaurants) • Personal and house hygiene (pharmaceuticals, watches) • Stationery(paper, toys, books, CDs) • House accessories(computers, dish-washer) • Services(flowers, animal food) • Clothing(cloth, leather) • Construction(cement, chemicals) • Cargo(general cargo)

  7. Models estimated • Discrete choice models • M1 - Multinomial logit model with linear effects(attributes linear and normalized) • M2 – Multinomial logit model with non-linear effects(effects coding) • M3 – Latent class model with linear effects(the same specification as in M1) • M4 – Latent class model with non-linear effects(the same specification as in M2) • Comparison between models through WTP measures(confidence intervals based on Delta method)

  8. Main results (I) • M1 – MNL, linear effect, attributes linear and normalized • Model fit: adj.Rho2 = 0.252 • Coefficients statistically significant, with the expected sign • Tariff plays the lion part in explaining preferences • SQ adversion

  9. Main results (II) • M2- MNLwith non-linear effects (effects coding) • Better fit (adj.Rho2 = 0.281) • All the reported coefficients are statistically significant • PLUBF is linear

  10. Main results (II) • M2 -MNLwithnon-linear effects (effects coding) • In line with prospect theory (Kahneman & Tversky, 1979)

  11. Main results (III) • M3 - LC with linear effects (same specification as in M1) Estimated latent class probabilities: Class1 = 0.50; Class 2 = 0.50 • Better fit (adj.Rho2 = 0.377) – almost equal class membership probabilities • Class 1 comprises more price-sensitive agents • Agents in Class 2 are more interested in LUB and PLUBF

  12. Main results (IV) • M4 - LC with non-linear effects (same specification as in M2 Estimated latent class probabilities: Class1 = 0.50; Class 2 = 0.5 Discriminant socio-economic variables to explain class membership (CART model): Number of customers (145) Number of deliveries per day(4,5) • Better fit (adj.Rho2 = 0.423), C1 price sensitive; C2 Bays sensitive • The same considerations of M3 apply here

  13. Main results (V) • WTP comparison P1 P2 P3 • Impact of Non-linear effects: M1 vs M2Overall efficiency loss for P1, P2, P3 = 18€, 49€, 6€ • Impact of Heterogeneity: M1 vs M3Overall efficiency loss for P1, P2, P3 = 198€, 396€, 481€ • Impact of joint Heterogeneity &Non-linear effects: M1 vs M4Overall efficiency loss for P1, P2, P3 = 435€, 614€, 693€

  14. Conclusion • Final remarks • The results obtained are relevant both from a theoretical as well as practical and policy-oriented perspective • The paper represents a first attempt at bridging the gap between theory, applied research and data needs • Relevant biases could characterize the results obtained if non-linearity & heterogeneity are not duly accounted for • There is a need for a sophisticated agent-specific model treatment to implement well-tailored and effective policies. • Future research • Similar investigation on retailers and own-account • Dealing with: i) interactive choice models; ii) Bayesian estimation methods; iii) sample size increment

  15. Thanks for your attention! • Questions? • Questions? • Questions? • Questions? • Questions?

More Related