Kenneth E. Train

Learn More
The simulation variance in the estimation of mixed logit parameters is found, in our application, to be lower with 100 Halton draws than with 1000 random draws. This finding confirms Bhat’s (1999a) results and implies significant reduction in run times for mixed logit estimation. Further investigation is needed to assure that the result is not quixotic or(More)
We develop a consumer-level model of vehicle choice to investigate the reasons behind the erosion of the U.S. automobile manufacturers’ market share during the past decade. Our model accounts for the influence of vehicle attributes, brand loyalty, product line characteristics, and dealerships on choice. We find that nearly all of the loss in market share(More)
We compare multinomial logit and mixed logit models for data on California households’ revealed and stated preferences for automobiles. The stated preference data elicited households’ preferences among gas, electric, methanol, and CNG vehicles with various attributes. The mixed logit models provide a much better fit to these data, and forecasting exercises(More)
We compare two approaches for estimating the distribution of consumers’ willingness to pay (WTP) in discrete choice models. The usual procedure is to estimate the distribution of the utility coefficients and then derive the distribution of WTP, which is the ratio of coefficients. The alternative is to estimate the distribution of WTP directly. We apply both(More)
This paper describes a recursive method for estimating random coefficient models. Starting with a trial value for the moments of the distribution of coefficients in the population, draws are taken and then weighted to represent draws from the conditional distribution for each sampled agent (i.e., conditional on the agent’s observed dependent variable.) The(More)