ABSTRACT
Maximum likelihood (ML) is by far the most popular method of estimation. It is generally
credited to R. A. FISHER (1890 – 1962), although its roots date back as far as J. H. LAM-
BERT (1728 – 1777), DANIEL BERNOULLI (1708 – 1782) and J. L. LAGRANGE (1736 –
1813) in the eighteenth century (see EDWARDS (1972) for a historical account). FISHER
introduced this method as an alternative to the method of moments (see Chapter 12) and to
the method of least squares (see the previous chapter). The former method FISHER criti-
cized for being arbitrary in the choice of the moment equations and the latter for not being
invariant under scale changes in the variables.
