[OPE-L:3058] Marxian Emprical Research

Andrew Trig (A.B.Trigg@open.ac.uk)
Thu, 19 Sep 1996 12:13:26 -0700 (PDT)

[ show plain text ]

From: Andrew Trigg
In reply to: Iwao Kitamura

In response to your two additional questions:

1. You are right that the key problem with this approach is to decide
which variables reflect quality (labour value) and which do not. I
have followed the strategy of replicating what the vast neoclassical
literature on wages does in this regard. For human capital theorists
such as Blaug age does provide an indicator of productivity. An older
worker has more experience, more human capital, and hence performs
more labour quality than a younger worker (other things being equal).
As workers get very old, of course, their productivity falls so a squared
variable is often put in which has a negative effect to temper the
positive effect on wages associated with age. As for sex, Becker argues
that a married man is more productive than a married woman, since he
has less distractions (!) and so interactive dummies, for male/married and
woman/married are included in the wage equation.

Your institutional objective may of course be correct. It may be because
old married men are the people in power that they pay themselves more
wages. Following this line of reasoning, however, leads to arguing that
no variations in wage are due to differences in quality. Professional
workers may earn more because they are the people who hold power; and
the same applies to people of high eductation - having a degree may
just be a screening device. All of this may be true. It may be that
we should just assume that all observed labour time is of the same
quality. In replicating the neoclassical approach to wage equations
the intention is merely to see what will happen to the resultant estimates
of labour values if ecuation/occupation/age do relate in some way to
labour quality. It would be better, of course, to have a rigorous
institutional/Marxian way of choosing which variables to use in this
decomposition, but since such a framework is not (I think) available then
a starting point is to run with the neoclassical approach to see what

2. Your second question about educational background is easier to
deal with. If two workers of the same occupation had different backgrounds
then this could be identified. The General Household Survey in the UK, for
example, gives information on what type of qualifications workers have -
e.g. ordinary level, advanced level, diploma, degree etc. I think this
information may also be available in the US panel which Patrick Mason
referred to (????) The way to think about this is that the variable in
the wage equation for, say, being a skilled worker, would have a certain
relationship with the wage, but at the same time the variable for
educational background would control for any differences in education
between workers of the same occupation.

Sorry that there have not been many/any strikes in Japan. Your time
will come, I'm sure.

Thanks again for all this response to my empirical work. It has in
fact kicked me into doing more work on it.

I would very much like to keep in touch with you and others interested
in this area.

In solidarity,