AN EXAMPLE SHOWING THAT A NEW TECHNIQUE FOR LAV ESTIMATION BREAKS DOWN IN CERTAIN CASES

An example is given that shows how, in certain situations, a recently proposed technique to approximate the least absolute value (LAV) estimates of the unknown parameters of a linear model will break down, even though a well defined, but not unique, solution to the standard LAV estimation problem exists. Even if breakdown does not occur, examples can be constructed that cast doubts on the ability of the above mentioned technique to approximate LAV estimates with full generality. © 1990.