Below five Matlab m files that can be used to estimate the consumer search model as described in “Maximum likelihood estimation of search costs” (with José Luis Moraga-González), European Economic Review, 52, 820-48, 2008.
The code is meant for academic research only. Users of this code (or a modified version of it) should reference the above paper. The code is provided without additional support. The code is tested using Matlab 7.8 (including optimization toolbox) on both Mac OS X and Windows XP.
The main file is:
maxlik.m - Reads the data and calls the other functions.
Accompanying functions are:
loglik.m - Computes the log-likelihood function;
f.m - Computes F(p) that solves the equilibrium condition;
condc.m - Computes the integrand used to calculate the critical search cost values;
conddc.m - Computes the derivative of the search cost values with respect to the parameters.
Price data used for the estimation:
pricedata.txt - Prices of Kingston KTT3614 memory chip.
Last updated: August 15, 2019