Subject: Re: A few here may have an opinion on this
From: Ian Lance Taylor <>
Date: 24 Oct 2002 09:27:18 -0700

"Wendel de Joode, Ruben" <> writes:

> I do not understand why a high IRR implies high productivity... can
> someone please explain this to me, as I feel that a high IRR could also
> imply that software companies simply charge too much money compared to
> the costs they have to incur to build the software? This means that
> customers end up paying (too) much money. The productivity of the
> software industry is not higher, instead, they are simply able to
> overcharge for their products.

Well, a standard economics argument is that if software companies do
indeed overcharge, then naturally people will see the opportunity to
build similar software and charge less for it.  Customers will choose
the lower-priced solution, and thus the new company will take over the
market, and make good money even if it is less than the original,
over-charging, company.

In other words, by the standard economic argument, it's unlikely for
there to be a long-term case of overcharging for a product.  This
implies that companies are charging approximately the right price, and
thus that high IRR does indeed imply high productivity.

But there are a lot of problems with that argument with respect to
software (e.g., very low marginal cost, and natural monopolies), so
let's ignore it.

Consider instead a product like Microsoft Word, which has come to
dominate its market segment.  Many companies developed word processing
products, and a lot of money was invested in their development.  This
is because there is a lot of value in the market.  However, we have
seen that it is the nature of the word processing market segment that
if different implementations are not interchangeable, a single
implementation will dominate.  Since Microsoft wrote the dominant
implementation, this permits them to capture the value of the entire
market segment.  In a sense, the price charged for Microsoft Word
reflects all the investment which was made in word processing
products.  It just happens that Microsoft is the only company which is
picking up the profit.  The other companies simply lost their

Anyhow, there is a simpler way of looking at the matter.  People
simply would not pay for software if it did not help them.  There is
no law saying you must buy a computer.  Therefore, the price charged
must be acceptable, even if the marginal cost to the seller is low.
This implies that software does indeed increase overall productivity.

> I do not see how this (IRR) leads to the conclusion that proprietary
> software is more productive for a country and leads to a higher GDP.
> Therefore I totally agree with Forrest to question why it is:
> 	"appropriate to start by looking at the IRR for software
> producers?"

I don't think you can use return on investment to come to any
conclusions about proprietary software versus free software.  It's
impossible to measure the productivity generated by free software.
Who can say how much proprietary software would never have been
written without the low startup costs permitted by free software?  Who
can say how much investment has been saved because free software sets
a lower limit on the functionality of basic software infrastructure?
Any answer to these questions now is no more than a wild guess.