Subject: Re: Open letter to those who believe in a right to free software
From: "Karsten M. Self" <>
Date: Tue, 26 Oct 1999 13:11:04 -0700 wrote:
> Stephen Turnbull wrote:

> >     Ben> Economics.  NOT software development.
> >
> > Yup.  Full house beats a flush.  Economists must practice learning
> > enough about other fields to make sufficiently accurate models of
> > those fields for economic purposes.  Software developers, except as a
> > hobby, do not do the reverse for economics AFAIK.  Seems like it would
> > generally be a waste of time.
> >
> And their success in modelling is measured by...other economists.

Economic models are also tested by making predictions about real world
behavior.  Succesful models tend to have better predictive power. 
Allegory:  John Maynard Keynes (father of Keynsian economics) got a lot
more credibility (after initally being laughed out of the house) by
making a killing in the stock market.  Granted, this was macro, not
micro, but....
> ie You know better.

> The usual assumptions appear to be somewhat distorted by the
> fact that:
>  - Vendor lock-in means that each software product is to some
>    extent a natural monopoly (lawyers originally were encouraged
>    to put Word Perfect to full use, they are still locked into it despite
>    years of pressure from clients to switch - other examples exist
>    if you are interested)

The term I think you're looking for is "non-substitutable good".  While
this may create a localized monopoly for a single user, this advantage
doesn't transfer to non-users (except through network effects, which you
weren't dicussing) in the way the price advantage of a natural monopoly
does, and it generally doesn't necessarily imply a market monopoly. 
Substitution is possible, but transition costs are high.

>  - The bulk of development costs lies in testing and debugging,
>    both of which are readily distributed

s/are/may be/

Depends greatly on the problem area and the software architecture. 
Complex, interdependent software is a bitch to debug in any
circumstances.  Distributing the task among many people with a shallow
understanding of the product might hurt rather than help.  This is among
the reasons that I argue (as does Stephen) for a high degree of
modularity in successful free software products.

>  - Software quality is frequently a market externality, indeed the
>    developer *benefits* from a certain level of bugs - it gives
>    consumers reasons to upgrade!

SWQ isn't a "market externality".  It can be used as a producer
advantage to manage lock-in and ensure forward migration, however, in a
unit-sales business model.  In this sense, it becomes a component of

Stephen's challenged the conclusion, but I think Ben's got a point,
though a limited one.  It's fairly well borne out by a number of
analysis of Microsoft, and even gets hinted at by the Bill itself in
some interviews.  

Note that even for proprietary software, this advantage disappears under
a services or leased-license business model in which customers are
entitled to use and run *any* version of the software.  I've seen more
than one product in whicht the vendor actively discouraged use of the
"latest and greatest" version of a product due to quality concerns.

>  - A significant portion of the value to a consumer lies not in the
>    product, but in guarantees of future support and development

This is true for any ongoing services type of business relationship,
such as law, advertising, or commercial mortgage backed securities
services.  As Stephen pointed out, this is merely another side of

>  - The incremental cost of distribution is approximately zero

As with any information good.  The issue's been dealt with in
economics.  No, it's not a characteristic of a "typical" good, but
software isn't absolutely unique in this way.
> > Of the list you gave, my opinion is that none violate "the usual
> > assumptions" about production.  This will need to be carefully
> > checked, of course.  Consumption of software does violate the usual
> > assumptions (it is non-rival and imperfectly excludable), but I am
> > considering those.
> >
> Please explain for each of the above properties why it does not
> violate "the usual assumptions"...

Ben, it would help us all if you could state what you think "the usual
assumptions" are.  I sense a disconnect.
> [...]
> >     Ben> and you appear to not see that software has some unusual
> >     Ben> properties as an economic good.
> >
> > Oh, indeed it has.  All goods are unique, of course.  But somehow the
> > same economic laws seem to hold over and over again.  That doesn't
> > prove anything; but experience shows that economists make better
> > models starting from the standard set of models rather than by
> > considering the exceptional aspects of each market.
> >
> A basic research question that I recommend to you.  To what extent
> is a proprietary program a natural monopoly?  In particular what are
> the real and perceived barriers in the market to switching products?

A natural monopoly, in the classic definition, arises where marginal
costs are falling over the entire range of production (output).  A
proprietary program is a natural monopoly only to the extent to which it
meets this definition.  The problem, with software, is that marginal
cost typically *isn't* falling -- it's constant (cost of cutting another
CD) or rises slightly (support and distribution costs experience
diminishing marginal returns).  Software offers typically huge
first-to-market benefits, but doesn't have monopoly benefits in the same
way that, say, wire-based distribution services (electricity and
telecoms) do.

If a rival good can be produced for lower cost or significantly higher
quality, and the switching costs are sufficiently low, then the
competitve advantages of the first proprietary program is lost (this
should sound familiar:  MS WinXX vs. Linux or xBSD).

For more information, see:

I'd strongly recommend that Ben take a look at a solid intermediate
micro text -- I've no doubt he could manage the material.  I recommend
Hal Varian, _Intermediate Microeconomics_
> (Certainly the industry belief is that it is a natural monopoly, and a

I don't think that this is the case.

The "industry belief" is that there is a monopoly.  However this
monopoly is based less on the factors typically attributed to natural
monopoly (diminishing costs of marginal production), and more on taking
advantage of lock-in, "tippy" demand, aggressive business practices, and
monopolizing (in the Sherman Antitrust Act sense) rather than monopoly
tendencies inherent in the market.  See _Information Rules_ (also by Hal
Varian: for expansion on some of these

Karsten M. Self (
    What part of "Gestalt" don't you understand?

SAS for Linux:
Mailing List:  body "subscribe sas-linux"