Abstract
Traditional multiple discrete–continuous (MDC) choice models impose tight linkages between consumers’ discrete choice and the continuous consumption decisions due to the use of a single utility parameter driving both the decision to choose and the extent of choice. Recently, Bhat (Trans Res Part B Methodol 110:261–279, 2018) proposed a flexible MDCEV model that employs a utility function with separate parameters to determine the discrete choice and continuous consumption values. However, the flexible MDCEV model assumes an independent and identically distributed (IID) error structure across the discrete and continuous baseline utilities. In this paper, we formulate a flexible non-IID multiple discrete–continuous probit (MDCP) model that employs a multivariate normal stochastic distribution to allow for a more general variance–covariance structure. In doing so, we revisit Bhat (Trans Res Part B: Methodol 109: 238-256, 2018) flexible utility functional form and highlight that the stochastic conditions he used to derive the likelihood function are not always consistent with utility maximization. We offer an alternate interpretation of the model as representing a two-step decision-making process, where the consumers first decide which goods to choose and then decide the extent of allocation to each good. We demonstrate an application of the proposed flexible MDCP model to analyze households’ expenditure patterns on their domestic tourism trips in India. Our results indicate that, if the analyst is willing to compromise on the strict utility-maximizing aspect of behavior, while also enriching the behavioral dimension through the relaxation of the tie between the discrete and continuous consumption decisions, the preferred model would be the flexible non-IID MDCP model. On the other hand, if the analyst wants the model to be strictly grounded on utility-maximizing behavior (which may also have benefits by way of welfare measure computations), and is willing to assume a very tight tie between the discrete and continuous consumption decision processes, the preferred model would be the non-IID traditional MDCP model.
Similar content being viewed by others
Notes
The baseline preference parameter of a good is the marginal utility at zero consumption of that good.
The discussion here pertains to the \(\gamma - {\text{profile}}\) utility form since it has been generally the case that the \(\gamma - {\text{profile}}\) comes out to be superior to the \(\alpha - {\text{profile}}\) function (see Bhat 2018b; Jian et al. 2017). However, the discussion in the rest of the paper, including the model formulation, is applicable for the \(\alpha - {\text{profile}}\) utility function as well.
\(\begin{gathered} {\text{At}}\,x_{k} = 0,\,\,\left. { \, \frac{{\partial U({\varvec{x}})}}{{\partial x_{k} }}} \right|_{{x_{k} = 0}} = \mathop {\lim }\limits_{{h \to 0^{ + } }} \,\left. {\frac{{U({\varvec{x}} + h) - U({\varvec{x}})}}{h}} \right|_{{x_{k} = 0}} \, = \,\mathop {\lim }\limits_{{h \to 0^{ + } }} \,\frac{{\left. {U({\varvec{x}} + h)} \right|_{{_{{x_{k} = 0}} }} - \left. {U({\varvec{x}})} \right|_{{_{{x_{k} = 0}} }} }}{h} = \mathop {\lim }\limits_{{h \to 0^{ + } }} \,\frac{{\left. {U({\varvec{x}} + h)} \right|_{{_{{x_{k} = 0}} }} - \left. {U({\varvec{x}})} \right|_{{_{{x_{k} = 0}} }} }}{h} \hfill \\ \,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\,\, = \mathop {\lim }\limits_{{h \to 0^{ + } }} \,\frac{{\psi_{kc} \gamma_{k} \ln \left( {\frac{h}{{\gamma_{k} }} + 1} \right) - 0}}{h} = \mathop {\lim }\limits_{{h \to 0^{ + } }} \,\frac{{\frac{\partial }{\partial h}\psi_{kc} \gamma_{k} \ln \left( {\frac{h}{{\gamma_{k} }} + 1} \right)}}{{\frac{\partial }{\partial h}h}} = \psi_{kc} .{\text{ Note:}}U({\varvec{x}} + h) = U(x_{1} ,x_{2} ,...,(x_{k} + h),...,x_{K} ). \hfill \\ {\text{At}}\,x_{k} > 0,\, \, \left. { \, \frac{{\partial U({\varvec{x}})}}{{\partial x_{k} }}} \right|_{{x_{k} = 0^{ + } }} = \mathop {\lim }\limits_{{x \to 0^{ + } }} \,\,\psi_{kc} \left( {\frac{{x_{k} }}{{\gamma_{k} }} + 1} \right)^{ - 1} = \psi_{kc} . \hfill \\ \end{gathered}\) Given the above, the \(\psi_{k\,d}\) parameters do not enter the marginal utility functions of Bhat’s (2018b) utility form, which as we discuss later, has implications for whether the formulation is always consistent with utility maximization.
Note also that the equality condition for the chosen goods, which is based on the C-preference parameters, automatically implies an inequality that \(\psi_{kc} - \lambda p_{k} > 0\). Such an inequality based on the C-preference parameters is not explicitly stated in the above conditions since it would be redundant.
In the special case that M = 0 (that is, none of the inside goods is consumed), Eq. (13) collapses to the following:
\(P\left( { \, {\varvec{x}}_{{}}^{*} } \right)\,\, = P(0,0,0,...0) = \,\,\,\,\displaystyle\int\limits_{{{\varvec{\eta}}_{D2} = - {\mathbf{\infty }}}}^{{{\varvec{\eta}}_{D2} = \tilde{\boldsymbol{V}}_{D2} }} {} \,{\varvec{f}}_{K - 1 + M} ({\varvec{\eta}}_{D2} ;\,{\mathbf{0}}_{K - 1 + M} ,{{\varvec{\Theta}}})\,d{\varvec{\eta}}_{D2} ,\)
which takes the form of a simple multivariate cumulative normal distribution function.
The symmetric nature of the multivariate normal density function is a distinct advantage over the asymmetric multivariate logistic density function used in Bhat (2018b). While the multivariate logistic has a closed form expression for the cumulative distribution function, computing the integral of the multivariate logistic density function with a combination of upper and lower limits cannot be collapsed to the evaluation of a single cumulative distribution function. However, while the multivariate normal cumulative distribution (MVNCD) function is not available in closed-form, the integral of the multivariate normal density function with a combination of upper and lower limits can be collapsed to the evaluation of a single MVNCD function. This is a particularly useful result for the proposed flexible MDCP model.
A small nuance is in order here. Bhat’s (2018b) flexible MDCEV uses the same variance for \(\tilde{\lambda }_{kk}\) and \(\omega_{kk}\), which leads to a closed-form likelihood expression. However, other than the distribution form, the IID D-C MDCP model is similar to the MDCEV in Bhat’s paper, because the scale does not matter in the discrete part of the IID D-C MDCP, and we can as well normalize \(\tilde{\lambda }_{kk} \,\) to \(0.5\mu^{2}\) (k = 1,2,…K). However, just to be consistent with the model development in Sect. "Revisiting Bhat’s (2018a, b) flexible MDC model structure", we will keep to the assumption that \(\tilde{\lambda }_{kk} \, = 0.5\), because it more clearly shows that the IID D-C MDCP model is a restricted version of the full flexible MDCP model proposed in Sect. "Revisiting Bhat’s (2018a, b) flexible MDC model structure".
References
Asgary, N., De Los Santos, G., Vincent, V., Davila, V.: The determinants of expenditures by Mexican visitors to the border cities of Texas. Tour. Econ. 3(4), 319–328 (1997)
Bhat, C.R.: A multiple discrete-continuous extreme value model: formulation and application to discretionary time-use decisions. Trans. Res. Part b: Methodol. 39, 679–707 (2005)
Bhat, C.R.: The Multiple discrete-continuous extreme value (MDCEV) model: role of utility function parameters, identification considerations, and model extensions. Trans. Res. Part b: Methodol. 42(3), 274–303 (2008)
Bhat, C.R.: New matrix-based methods for the analytic evaluation of the multivariate cumulative normal distribution function. Trans. Res. Part b: Methodol. 109, 238–256 (2018a)
Bhat, C.R.: A new flexible multiple discrete-Continuous extreme value (MDCEV) choice model. Trans. Res. Part b: Methodol. 2018(110), 261–279 (2018b)
Brida, J.G., Scuderi, R.: Determinants of tourist expenditure: a review of microeconometric models. Tourism Management Perspectives 6, 28–40 (2013)
Calastri, C., Hess, S., Daly, A., Maness, M., Kowald, M., Axhausen, K.: Modelling contact mode and frequency of interactions with social network members using the multiple discrete-continuous extreme value model. Transp. Res. Part C 76, 16–34 (2017)
Enam, A., Konduri, K.C., Pinjari, A.R., Eluru, N.: An Integrated choice and latent variable model for multiple discrete-continuous choice Kernels: application exploring the association between day level moods and discretionary activity engagement choices. J. Choice Modell. 26, 80–100 (2018)
Hanemann, W.M.: Discrete/continuous models of consumer demand. Econometrica 52(3), 541–561 (1984)
Hendel, I.: Estimating multiple-discrete choice models: an application to computerization returns. Rev. Econ. Stud. 66(2), 423–446 (1999)
Herter, M.M., Pizzutti, C., Pinto, D.C.: ‘Man, I Shop like a Woman!’ the effects of gender and emotions on consumer shopping behaviour outcomes. Int. J. Retail Distrib. Manag. 42(9), 780–804 (2014)
Jian, S., Rashidi, T.H., Dixit, V.: An analysis of carsharing vehicle choice and utilization patterns using multiple discrete-continuous extreme value (MDCEV) models. Trans. Res. Part a: Policy Pract. 103, 362–376 (2017)
Khan, M., Machemehl, R.: Analyzing tour chaining patterns of Urban commercial vehicles. Trans. Res. Part a: Policy Pract. 102, 84–97 (2017)
Kim, J., Allenby, G.M., Rossi, P.E.: Modeling consumer demand for variety. Mark. Sci. 21(3), 229–250 (2002)
Mulky, A.G.: Distribution challenges and workable solutions. IIMB Manag. Rev. 25(3), 179–195 (2013)
National Sample Survey Office (NSSO). (2016): Domestic tourism expenditure, NSS 72nd round (July 2014-June 2015). Ministry of Statistics and Programme Implementations (MOSPI), Government of India. http://microdata.gov.in/nada43/index.php/catalog/138/. Accessed in June 2020
Nicolau, J.L., Mas, F.J.: Stochastic modeling: a three-stage tourist choice process. Ann. Tour. Res. 32(1), 49–69 (2005)
Pellegrini, A., Sarman, I., Maggi, R.: Understanding Tourists’ expenditure patterns: a stochastic frontier approach within the framework of multiple discrete-continuous choices. Transportation 48, 931–951 (2021)
Pinjari, A.R.: Generalized extreme value (GEV)-based error structures for multiple discrete-continuous choice models. Trans. Res. Part b: Methodol. 45(3), 474–489 (2011)
Pinjari, A.R., Bhat, C.R.: A multiple discrete-continuous nested extreme value (MDCNEV) model: formulation and application to non-worker activity time-use and timing behavior on weekdays. Trans. Res. Part b: Methodol. 44(4), 562–583 (2009)
Pollak, R.A., Wales, T.J.: Price-augmenting returns to scale: an application to nonseparable two-stage technologies. Rev. Econ. Stat. 74(2), 213–220 (1992)
Satomura, T., Kim, J., Allenby, G.M.: Multiple-constraint choice models with corner and interior solutions. Market. Sci. 30(3), 481–490 (2011). https://doi.org/10.1287/mksc.1110.0636
Saxena, S., Pinjari, A.R., Bhat, C.R.: Multiple discrete-continuous choice models with additively separable utility functions and linear utility on outside good: model properties and characterization of demand functions. Trans. Res. Part B 155, 526–557 (2022)
Sobhani, A., Eluru, N., Faghih-Imani, A.: A latent segmentation based multiple discrete continuous extreme value model. Trans. Res. Part b: Methodol. 58, 154–169 (2013)
Venugopal, P.: Urban orientation of rural consumers: implication for consumer goods distribution. Int. J. Rural. Manag. 8(1–2), 107–119 (2012)
von Haefen, R.H., Phaneuf, D.J.: Estimating preferences for outdoor recreation: a comparison of continuous and count data demand system frameworks. J. Environ. Econ. Manag. 45(3), 612–630 (2003)
von Haefen, R.H., Phaneuf, D.J.: Kuhn-Tucker demand system approaches to non-market valuation. In: Scarpa, R., Alberini, A. (eds.) Applications of simulation methods in environmental and resource economics, pp. 135–157. Springer Netherlands, Dordrecht (2005). https://doi.org/10.1007/1-4020-3684-1_8
Wafa, Z., Bhat, C.R., Pendyala, R.M., Garikapati, V.M.: Latent-segmentation-based approach to investigating spatial transferability of activity-travel models. Transp. Res. Rec. 2493(1), 136–144 (2015)
Wales, T.J., Woodland, A.D.: Estimation of consumer demand systems with binding non-negativity constraints. J. Econ. 21(3), 263–285 (1983)
Wang, Y., Davidson, M.C.: A review of micro-analyses of tourist expenditure. Curr. Issue Tour. 13(6), 507–524 (2010)
Yang, C.W., Wu, C.L., Lu, J.L.: Exploring the interdependency and determinants of tourism participation, expenditure, and duration: an analysis of Taiwanese citizens traveling abroad. Tour. Econ. 27(4), 649–669 (2021)
Yonezawa, K., Richards, T.: Consumer risk-reduction behavior and new product purchases. Manag. Decis. Econ. 38, 1003–1016 (2017)
Zemke, D.M.V., Jay, N., Stowe, S., Katie, K.: Hotel cleanliness: Will guests pay for enhanced disinfection? Int. J. Contemp. Hosp. Manag. 27(4), 690–710 (2015). https://doi.org/10.1108/IJCHM-01-2014-0020
Acknowledgements
This research was undertaken as part of a SPARC project funded by the Indian Ministry of Education for encouraging international collaborations. Empirical data used in this study were collected by the National Sample Survey Office of the Indian Ministry of Statistics and Program Implementation. The authors are grateful for the support provided by both ministries. The second author would like to thank Tarun Rambha for helpful discussions on non-linear optimization.
Author information
Authors and Affiliations
Contributions
Author contributions to the paper are as follows: study conception and design: CB, SS, AP; analysis and interpretation of results: SS, AP, CB, AM; draft manuscript preparation: SS, AP, CB, AM. All authors reviewed the results and approved the final manuscript.
Corresponding author
Ethics declarations
Conflict of interest
On behalf of all authors, the corresponding author states that there is no known conflict of interest.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
About this article
Cite this article
Saxena, S., Pinjari, A.R., Bhat, C.R. et al. A flexible multiple discrete–continuous probit (MDCP) model: application to analysis of expenditure patterns of domestic tourists in India. Transportation (2023). https://doi.org/10.1007/s11116-022-10364-y
Accepted:
Published:
DOI: https://doi.org/10.1007/s11116-022-10364-y