Review of Industrial Organization 18: 273–274, 2001.
© 2001 Kluwer Academic Publishers. Printed in the Netherlands.
A Monopolist Would Still Charge More for
Windows: A Comment on Werden’s Reply
BERNARD REDDY, DAVID EVANS, ALBERT NICHOLS and RICHARD
NERA, 1 Main Street, Cambridge, MA 02142, U.S.A.
Abstract. Werden’s reply to our comment does nothing to cast doubt on our fundamental conclu-
sion: Microsoft’s pricing of Windows is inconsistent with the government’s claim that Windows is a
monopoly, protected by high entry barriers.
In his reply to our comment (Reddy et al., 2001), Werden (2001) denies he had
even attempted to show that Microsoft’s actual price for Windows could have been
selected by a proﬁt-maximizing Windows monopolist. He claims to have shown
only that the heterogeneity of PC demand “may go a long way toward explaining
why the price of Windows might be as low as it is”. But he has not done this.
Werden has shown by example that heterogeneity could be important in theory.
But as we demonstrated in our comment, because his examples rest on a set of
incorrect or implausible assumptions about parameter values, they shed no light on
actual Windows pricing. In his reply, Werden makes no serious effort to refute our
In our comment, we used semi-log demand equations and parameters consistent
with real-world data to demonstrate that heterogeneity most likely has a negligible
effect on the monopoly price. In his reply, Werden says that our two segments
are not heterogeneous enough. To deal with this criticism, we tested other para-
meter values that preserved the mean PC price observed in 1999 ($1,600) and
the assumed (high) overall elasticity of −2, but widened the differences between
groups. We found that the proﬁt-maximizing prices almost always increased and
never fell appreciably. For example, with initial PC prices at $500 and $2,333
and elasticities of −3.5 and −1 for the low- and high-priced groups, respectively,
the proﬁt-maximizing price more than tripled from the examples in our original
Werden notes, correctly, that current upgrade sales are complementary to past
Windows sales, so a detailed analysis of complementary sales would discount
expected future complementary sales to the present. In his critique, however, he
adjusts only for inﬂation and does not discount future sales. This issue is minor
in any case: using Werden’s ﬁgure of $150 for complementary sales in the ex-