The need for speed: all-in-one versus iterative-shortcut approaches for solving general equilibrium models of urban transportation and land use

General Equilibrium models of Land-Use and Transport (to which we will hereafter refer to as GELUT), as conceived and developed in the contributions of Anas and Kim (1996), Anas and Xu, (1997), Anas and Liu (2007) and applied empirically by Anas and Hiramatsu (2012; 2013) are becoming established as a widely accepted framework to model the urban economic environment. The approach provides a consistent quantitative method to perform welfare analyses of transport investments, land-use regulations, externality-pricing policies and reforms of a local tax system. Such urban general equilibrium models have become recognized as alternatives to spatial interaction/gravity models (Mackett, 1991), Land Use and Transport Interaction (LUTI) models (see for instance TIGRIS-XL by Zondag and de Jong, 2011 to mention only one) and disequilibrium (i.e. adjustment-based) micro-simulation approaches such as UrbanSim (Waddell, 1998c, 2000, 2002; Waddell et al., 2003;). 74 The latter approaches deviate from a general equilibrium approach in many ways. Some of them treat production and household income as exogenous. Others do not depend on endogenously derived demand and supply functions and may rely instead on ad hoc price setting mechanisms. Equipped with these mechanisms, LUTI models are intrinsically capable of replicating the observed correlation between many important variables of the urban economy (e.g. residential and employment density). For this reason, they are (in general) characterized by larger forecasting power compared to fully microfounded approaches. However, the eclectic sampling of behavioral mechanisms from the systematic and logical panoply of microeconomic theory translates into a reduced capacity to perform reliable welfare analysis. Because of this state-of-affairs, a wide gap (that includes too many aspects to cover meaningfully in one chapter) exists between GELUT models and the above approaches. Many years after the conception of the GELUT approach, its true possibilities and limitations have not yet been fully explored. Knowledge over their properties and mechanics accumulates through sporadic applications, or from the closely related field of Computable