Monte Carlo estimation of the mixed logit model using low-discrepancy sequences

Monte Carlo estimation evaluates a function at different points to compute an approximated average. Traditionally this is done at random points. Quasi-random sequences are strategically constructed deterministic points that aim to be more efficient than random points. The golden ratio has unique properties to construct low-discrepancy sequences that are compared to traditional quasi-random methods. The methods are compared by estimating the mixed multinomial logit model for panel data.