Competition and substitution between public transport modes

Review and meta-analysis of inter-modal cross-elasticity evidence

“…Notably though, the noise valuation meta-analysis of Bristow et al (2015) achieved a figure of 0.86 even though these are intrinsically variable valuation estimates. The meta-analysis of cross-elasticities between public transport modes reported in the predecessor paper of Fearnley et al (2018) recovered an adjusted R 2 of 0.43. Studies that specified linear-additive functions, where the goodness of fit is not directly comparable, achieved adjusted R 2 s of between 0.12 and 0.34 (Hensher, 2008), 0.27 (Kremers et al, 2002, between 0.22 and 0.68 (Holmgren, 2007), 0.43 (Brons et al, 2002) and between 0.28 and 0.34 (Espey, 1998).…”

“…In support of this, the Wardman (2014) meta-analysis of own-price elasticities found SP choice based elasticities to be around twice those of equivalent RP values whilst the Wardman (2012) meta-analysis of time based own-elasticities found the ratio to be in the range 25% to 70% larger, although Kremers et al (2002) and Hensher (2008) are less clear-cut in this regard. In the specific context of cross-elasticities, the Fearnley et al (2018) meta-analysis reports SP crosselasticities to be twice the RP equivalents.…”

“…We therefore consider it impressive that we have been able to recover such a large number of statistically significant effects which generally seem credible. In part, the large data set of 1096 observations will have helped, extending considerably on the 171 in the cross-elasticity metaanalysis reported by Fearnley et al (2018). Our sample compares favourably with the 444 observations for one of the first meta-analyses of value of time (Wardman, 1998) 9 , the 258 for the first meta-analysis of noise valuations (Bristow et al, 2015), and the samples of 1633 and 427 for the most extensive meta-analyses of price elasticities (Wardman, 2014) and time based elasticities (Wardman, 2012) respectively.…”

“…These are all dated and tend to cover relatively few observations and a limited set of cross-elasticity types. More recent reviews are provided by Fearnley et al (2017Fearnley et al ( , 2018, covering a larger number of studies and a broader range of cross-elasticity terms. We here build upon the latter reviews with considerably more data and greater detail surrounding variables and modes to report the most extensive review of inter-modal cross-elasticities yet undertaken.…”

“…We here build upon the latter reviews with considerably more data and greater detail surrounding variables and modes to report the most extensive review of inter-modal cross-elasticities yet undertaken. Whilst Fearnley et al (2018) is novel in reporting a meta-regression model on cross-elasticity evidence 1 , it is restricted to public transport modes and a far smaller data set than here, and the meta-analysis reported in this paper provides insights from a far larger range of explanatory variables and is the first ever that covers all modes and all attributes.…”