[DMCA-Activists] CEPR: Opening Doors and Smashing Windows

[Seth Johnson]

(PDF text pasted below.  -- Seth)

-------- Original Message --------
Subject: [Upd-discuss] Opening Doors and Smashing Windows:
Alternative  Measures for FundingSoftware Development
Date: Mon, 24 Oct 2005 17:14:49 -0400
From: robert weissman <address@hidden>
To: address@hidden

The Center for Economic and Policy Research has just released the
paper, "Opening Doors and Smashing Windows: Alternative Measures
for Funding Software Development" by Dean Baker, CEPR economist
and Co-Director.

This paper examines the ways in which copyrights and patents
(intellectual property rights) lead to inefficiency and waste,
and outlines alternative mechanisms for supporting software
development.

The paper can be accessed at:
http://www.cepr.net/publications/windows_2005_10.pdf.

Kathryn Bogel
E-mail: address@hidden
Center for Economic and Policy Research
1611 Connecticut Ave NW, Suite 400
Washington, DC 20009
Phone (202) 293-5380 x111
Fax (202) 588-1356
_www.cepr.net <http://www.cepr.net/>_

_______________________________________________
Upd-discuss mailing list
address@hidden
http://lists.essential.org/mailman/listinfo/upd-discuss

---

Opening Doors and Smashing Windows: 

Alternative Measures for Funding Software Development 

By: Dean Baker 

Center for Economic and Policy Research 
1611 Connecticut Avenue, NW 
Suite 400 
Washington, D.C. 20009 

Tel: 202­293­5380 
Fax: 202­588­1356 

www.cepr.net 

October 2005 

Table of Contents 

 Executive Summary 
 Introduction 
 The Inefficiency of Copyright and Patent Protection in the
Software Industry 
 Deadweight Loss: Consumers Cannot Buy Software at the Cost of
Production 
 IPR Protection ­­ Impeding Innovation and Promoting Duplication 
 Rent Seeking Activities of IPR Holders 
 21st Century Alternatives to Copyrights and Patents 
 The Competition Between IPRs and Alternative Mechanisms 
 Conclusion 
 Appendix 

About the Authors 

Acknowledgments 

Dean Baker is Co­Director of the Center for Economic and Policy
Research.

John Schmitt provided helpful commentson an earlier version of
this paper.



Executive Summary 

Copyrights and patents are forms of government intervention in
the market that are relics of the medieval guild system. They are
an outdated and inefficient means to support creative and
innovative work in the 21 st century. These government­granted
monopolies lead protected software to sell at prices that are far
above the free­market price. In most cases, in the absence of
copyright and patent protection, software would be available over
the Internet at zero cost.

This paper examines the ways in which copyrights and patents
(intellectual property rights, or IPRs) lead to inefficiency and
waste, and outlines alternative mechanisms for supporting
software development. It shows that: 

  - Copyright and patent protection may impose costs on consumers
of between $80 and $120 billion a year, compared to a situation
in which all software was available at its competitive­ market
price. The pure efficiency loss to the economy could be in the
neighborhood of $70 to $110 billion a year. These losses dwarf
estimates of the losses from other forms of protectionism, such
as tariffs or quotas on imported goods; 

  - A substantial portion of the resources devoted to software
development are currently wasted due to IPRs. This is a result of
the fact that IPR protection leads to unnecessary duplication, as
developers have substantial incentive to produce software that
simply replicates the function of existing software. In the
absence of IPR protection, developers could better spend their
time improving existing software. IPRs also provide incentives
for software locks and secrecy, which impede the process of
software development. It is likely that a substantial portion of
software development (possibly a majority) is misdirected as a
result of the market distortions created by IPRs; 

  - IPRs also lead to large amounts of waste by providing
incentives for rent­seeking activity.

This waste includes expenditures for advertising and marketing,
and payments to lawyers and lobbyists. Those enjoying IPR
protection have also been able to impose costs on third parties,
for example by requiring Internet Service Providers to monitor
activities or universities to take steps to reduce the amount of
unauthorized copying of IPR protected material on their premises.
IPR holders have also secured laws that restrict the development
of software and hardware designed to support better searches and
digital reproductions; 

  - There are feasible alternative mechanisms for supporting
software development. One mechanism outlined in the paper would
create a "Software Development Corps," which would be a series of
competing government funded software corporations. An annual
appropriation of $2.1 billion (approximately 0.08 percent of
federal spending) should be enough to support the work of
approximately 20,000 software developers. The government should
be able to recoup most, if not all, of this money through the
lower price it will pay on the computers and software it
purchases. The remaining benefit would be the equivalent of a tax
cut to consumers in the range of $80 to $120 billion a year. This
money would provide a substantial stimulus to the economy and
lead to the creation of millions of jobs.

A unique feature of this proposal for a Software Development
Corps is that it could exist side­by­side with the existing
copyright and patent system. Software developers, such as
Microsoft, could continue to produce IPR­protected software and
sell it in the market. If they actually produce software that is
sufficiently superior to justify paying the IPR­ protected price,
then consumers will still purchase it. In other words, the
alternative mechanisms described in this paper provide a basis
for a market test of the relative efficiency of IPR­supported
research. Those who believe that IPRs are the most efficient
mechanism for supporting software development should favor this
sort of test, which would prove their case.


Introduction 

The computer software industry is usually considered to be at the
cutting edge of U.S. technology. At the same time, the industry
relies largely on government protectionism in the form of
copyrights and patents, relics of the feudal guild system, to
finance the development of new software. The reliance on these
antiquated mechanisms leads to both economic inefficiency for the
economy as a whole, and poorer quality software than would be
possible in an environment in which all software was placed in
the public domain.

This paper outlines the ways in which patents and copyrights in
software lead to economic inefficiency.

It also presents alternative mechanisms that could be used to
finance the software development that is currently supported by
copyright or patent protection. Specifically, it suggests that a
mix of a system of direct government funding for software
development and a system of individual vouchers could be a more
efficient mechanism for financing the development of new
software. All the new software developed under both systems would
be placed in the public domain so that it could be used at zero
cost and freely modified by other developers. 1 


The Inefficiency of Copyright and Patent Protection in the
Software Industry 

One of the most basic principles in economics is that efficiency
is maximized when products sell at their marginal cost of
production. Copyright and patent protection is explicitly
designed to prevent marginal­cost pricing by providing copyright
and patent holders with a government­enforced monopoly on the
protected product. This monopoly allows the holder of these
intellectual property rights (IPR) to charge prices that are far
above the marginal cost of production, which would be the
expected equilibrium price in a freely competitive market.

Of course, this inefficiency may make economic sense if the
monopoly provides incentives to innovate.

However, the overriding policy question is whether copyright and
patent monopolies are the best way to provide incentives for
software development. To make this assessment, it is necessary to
determine the costs of IPRs in software and compare them to the
cost of other mechanisms for financing software development.

There are three distinct ways in which IPRs in software lead to
economic inefficiency:

1) The gap between the IPR­protected price and the
competitive­market price (which would be zero for most software,
since it can be transferred costlessly over the Internet) leads
to a deadweight efficiency loss. There are many consumers who
would be willing to pay at least the marginal cost of producing
another copy of the software, but not the IPR­protected price.
[This is exactly the same argument that is made for the benefit
of free trade, although the inefficiencies that result from most
tariffs, which are in the range of 10­ 20 percent, are trivial
compared to the inefficiencies that result from IPR protection.] 

2) The fact that certain types of software are proprietary, and
cannot be freely adapted to develop new and better software,
impedes the process of software development. Research proceeds
most rapidly when all findings are publicly available and can be
freely shared by the community of researchers. The fact that some
software is proprietary also leads to duplication of effort,
which would serve no purpose, if all software were freely
available.

3) Holders of IPRs in software devote resources to protecting
their IPRs, and also force others to devote resources to protect
their IPRs. Specifically, software producers spend resources
designing locks, contracting services to monitor usage of their
software, hiring lobbyists and contributing to politicians to
protect their IPRs, hiring lawyers to enforce their IPRs, and
demanding that the government and even third parties such as
Internet Service Providers and universities take measures to
protect their IPRs.

It is worth examining in somewhat more detail the nature and size
of the economic costs associated with each of these sources of
inefficiency before examining the feasibility of alternative
methods of supporting software innovation.


Deadweight Loss: Consumers Cannot Buy Software at the Cost of
Production 

A vast body of economic research attempts to measure the size of
the loss to consumers that results from tariffs or other taxes
that create a gap between what it costs sellers to produce a
product and the price that consumers have to pay in the market.
Economists argue that the gap between the two should be as little
as possible, and ideally zero, because this gap is a source of
pure waste.

The logic here is simple. Suppose a foreign apparel manufacturer
can produce a shirt for $20, counting all costs,
includingshippingandanormalprofit. If there is a 10 percent
tariff then the shirt will sell in the United States for $22.
This means that consumers whowouldhavebeenwilling to pay more
than $20, but less than $22 will not buy the shirt, because of
the tariff. This is the loss due to trade protection that leads
most economists to oppose tariffs. This loss is a pure loss to
the economy -- in contrast to the extra $2 paid by people who
actually purchase the shirt. In that case, the $2 lost by
consumers is transferred to the government. The governmentcan
thenuse this money for some other purpose.

However, there is no money transferred when people don't buy the
shirt because of the tariff. The people who end up not buying the
shirt are simply madeworse off because of the tariff.

Figure 1 
Deadweight Loss From $2 Tariff on Shirts 

Identical logic applies to the IPR­protected price of software,
except in the case of software, most products would be available
at little or no cost in the absence of IPR protection. Instead of
raising the price by 10 percent, as might be the case with a
typical tariff, with software, the only reason that there is a
price at all is because of the IPR protection. Various software
applications (e.g. spreadsheets, word processing programs,
graphic programs) that would be available for free over the
Internet, can often be sold for several hundred dollars, because
the government grants a monopoly to the copyright or patent
holder. People who try to distribute this software for free face
fines, and possibly even criminal charges by the government.

It is possible to use the same basic graph to show the losses
from IPR protection with software, but the numbers would be much
larger, as shown in the graph below.

Figure 2 
Deadweight Loss From Zero Marginal Cost Software, Selling at $100 

Since the gap between the cost of production of software and the
IPR­protected price is so much larger, the number of consumers
who end up not buying IPR protected software is likely to be much
larger, and the potential loss to consumers will be much larger
as well.

The logic here is straightforward. If a tariff raises the price
of a shirt by $2, then the most benefit that any consumer could
have lost because of this tariff is just less than $2.00 -- say
$1.99. If the shirt was in fact worth $2.00 more to the consumer
than the competitive market price (i.e. $22) then she would have
bought the shirt.

There is only an issue if the shirt was worth more than $20 to
the consumer, but less than $22 -- in which case she would have
bought it at the competitive market price, but not at the price
charged in the market with a tariff.

In the case of IPR protection of software, there are likely to be
many people who might have found the software useful and would
have even been willing to pay some price to use it, but less than
the $100 IPR­protected price shown in the graph. In this case,
the maximum loss to consumers would be just less than the IPR
protected price (i.e. up to $100), since that is the most that
someone would have been willing to pay for the software, who did
not actually buy it.

It is possible to do some rough calculations to get a general
estimate of the potential gains to consumers, and efficiency
gains to the economy, from eliminating IPR protection in
software. Table 1 shows projections of gains from eliminating IPR
protection for software in the market for notebook computers, the
market for desktop computers, and the market for computer and
video games. (For a fuller explanation of these calculations see
the appendix.) The projections assume alternatively that demand
is relatively unresponsive to price (low elasticity), moderately
responsive to price, and highly responsive to price.

Table 1 
Gains to Consumers from Eliminating IPRProtection for Computer
Software 

It is assumed that the price reductions are very large in all
cases, which leads to substantial gains to both consumers and to
the economy in the form of increased output. The calculations in
the table assume that the average sale price of both a notebook
and desktop computer would fall by $200 (including installed
software applications), if all software were available at zero
cost. They also assume that computer purchasers would use an
average of $600 of additional software (at current market prices)
if all software were available at zero cost. The calculations
assume that all video and computer games would be available at
zero cost over the Internet.

Table 2 summarizes the gains to consumers and the net efficiency
gains (after deducting losses to producers) in the three
scenarios.

Table 2 
Gains to Consumers and net Gains to the Economy FromEliminating
IPRs in Software 

The projected gains in all three cases are extremely large. In
the lowelasticity case, the gain to consumers is projected as
$84.2 billion a year, an amount that is more than 50 percent
larger than what the federal government spends on roads each
year. The projected gains to consumers in the high elasticity
case are $115.5 billion, more than twice theannual federal
transportationbudget.Theprojectedgains to the economy are
similarly large. The net surplus in the low elasticity case is
$74.1 billion a year, an amount that exceeds the
GDPofGuatemala.Thenet surplus in the high elasticity case is more
than $100 billion a year, an amount that is equal to
approximately 0.8 percent of U.S. GDP. Very few economic policies
yield these sorts of potential efficiencygains.

Thereare two important points to noteon these calculations, in
addition to the fact that they are extremely crude projections.
First, they are based on worldwide estimates of the computer
market. Much of these projected gains would accrue to people
outside of the United States. While there is nothing wrong with
this, presumably it would be desirable to devise a method to
share the cost of developing software, at least among middle
income and wealthy countries, whocanafford topaypartof this
burden.Devising a system to share the costs of an alternative
system for financingsoftwaredevelopment
internationallywouldrequire some coordination. However, would
likely be far simpler and easier to enforce than the complex
system currently in place to standardize and enforce IPRs
internationally.

Thesecondpoint is that thesecalculations onlyassumegains
associated with the purchase of new computers. If software were
freely available over theweb, tensofmillionsofcurrent owners of
computers would take advantage of the opportunity to download
better software at zero cost. This would lead to a large one­time
gain (probably several times larger than the annual gain assumed
in these calculations) that is not captured in these projections.
This one­time gain in consumer surplus could easily run into the
hundreds of billionsofdollars, if
severalhundredmillioncomputerowners download software worth an
average of several hundred dollars per computer.


IPR Protection -- Impeding Innovation and Promoting Duplication 

The second major type of inefficiency associated with IPR
protections in software are the ways in which IPR protection
impedes the development of new software and leads to wasteful
duplication. The first point is easy to describe, although
difficult to quantify. If we imagine the scientific process as
being a collective exercise in which everyone builds on each
other's progress, and learns from mistakes, then the process will
obviously advance most quickly in an environment in which
information flows most freely. If software developers have ready
access to the work of other developers, and can freely integrate
this work into their own projects, then the development process
will proceed much more quickly than if developers are forced to
work in isolation or cannot borrow from the discoveries of
others.

IPR protection in software leads to both sorts of problems.
Companies seeking to profit from software copyrights or patents
will not make information on their progress publicly available
until they actually file for IPR protection. This is simply the
nature of the incentives provided by the IPR system. If a company
allowed information to get out before it had filed for IPR
protection, then a competitor could use this information to get a
copyright or patent of their own. Even when a company does file
for IPR protection, it will typically only disclose the
information needed to gain protection -- it will still keep much
of its work secret. Software producers are likely to continue to
maintain as much secrecy as possible around their process of
software development, both to prevent aiding competitors and to
preserve an edge in developing future products.

Of course, even when a product is available, competitors cannot
freely build on it if the product is subject to IPR protection,.
In other words, a software designer does not have the option to
develop an extension of an IPR­ protected piece of software,
which might make it more useful, without the permission of the
IPR holder. This could prevent many useful adaptations of
software, since it can be costly and time consuming to negotiate
an arrangement with an IPR holder. In many cases, software
developers may simply choose to direct their energies elsewhere.

There can also be instances where ambiguity about the nature of a
software copyright or patent, by itself, inhibits progress. For
example, if Microsoft establishes a reputation for aggressively
defending its IPRs, it may discourage software developers from
pursuing lines of research that are not actually infringing on
Microsoft's IPRs, but which may still be close enough to prompt
lawsuits. The asymmetry of resources between Microsoft and a
small software developer is sufficiently large that many
developers may simply decide to abandon a path of research
altogether rather than get into a long legal battle over IPRs.

The existence of IPR protection in software also provides
incentives for wasteful duplication. If there is an application
subject to IPR protection that proves to be popular (e.g. a
word­processing program or spreadsheet program), there is a
powerful incentive for other developers to produce a competitive
program that does not infringe on the copyright or patent. While
there is nothing wrong with giving consumers a choice of
programs, in many cases the sole motivation may be that the
initial program was subject to IPR protection. In other words, if
a popular spreadsheet program were available at zero cost over
the web to anyone who wanted to use it, there would be little
incentive for someone to attempt to develop an alternative, even
if the alternative would be subject to IPR protection. If the
original program lacked important features, the most efficient
action for a developer to do would be to modify the existing
program to incorporate the new feature, rather than create a
near­duplicate of the original program ---all the while being
careful not to infringe on the original developer's IPRs---
simply to add an additional feature. However, if the original
program is subject to IPR protection, then an alternative can
potentially take away a substantial portion of the monopoly
profits earned by the first program.

This sort of competition is actually desirable in a world where
programs sell for high prices due to IPR protection, since
competition will lead to lower prices for consumers. However, in
a world without IPR protection, there would be little reason to
design a second program that essentially just replicates the
tasks already performed by an existing program. In most cases,
the time of software designers could be much more productively
spent improving the original program, or on other projects.

Unfortunately, it is difficult to estimate the losses due to
secrecy in development or unnecessary duplication. In the
pharmaceutical industry, approximately two thirds of research
spending by the industry goes to develop duplicative rather than
breakthrough drugs. 2 While there is some value associated with
such copycat drugs (some patients may have bad reactions to one
drug, but not to the copycat drug) this research is almost
certainly less valuable than research on a breakthrough drug. It
is probably reasonable to assume that somewhere in the
neighborhood of half of the patent supported drug research is
wasted on such duplicative research.

There is no obvious basis for projecting the extent to which
software development would be advanced if all research was public
and freely available to other developers. While the availability
of more information would surely lead to better software that is
developed more quickly, it is really only possible to guess at
the extent of the gains from increased openness.

One further complication in projecting the gains in this respect
from eliminating IPR protection is that it is not even clear how
much software development currently depends on IPRs. While the
Commerce Department provides estimates of spending on software
(approximately $190 billion in the United States in 2005), much
of this spending is for custom software programs that are
tailored for the needs of a specific user (such as
customer­service software for a commercial bank, rental­ car
agency, or online retail store). While the software developers
may obtain IPRs for the software they install, most of the fees
involved are for the design and installation of a firm­specific
system. The lack of IPR protection would probably have little
impact on the price of these systems or the ability of these
software developers to charge for their work.

A crude way of projecting the amount of software development that
depends on IPR protection is to use Microsoft's repor ted
expenditures as a basis of calculation. In its quarterly
statement for the second half of 2004, Microsoft reported that it
spent $3.0 billion on research and development in the second half
of 2004. 3 Assuming that this it spent roughly the same amount in
the first half, this would mean that Microsoft spent a total of
$6 billion on software development in 2004. If this sum is
doubled to include the spending of other software developers,
then it implies that the industry spent $12 billion of research
supported by IPR protection in 2004.

Using this number as a starting point, it is possible to project
the amount of wasted expenditures in software development due to
IPR protection. Table 3 shows projections for low, middle, and
high waste scenarios, making alternative assumptions about the
waste due to unnecessary duplication and secrecy. In the case of
unnecessary duplication, the low, middle, and high waste
scenarios assume alternatively that 20 percent, 40 percent, and
60 percent of expenditures are devoted to duplicative activity
that would not be carried through in the absence of IPR
protection. The losses due to secrecy are assumed to be
alternatively 10 percent, 20 percent, and 30 percent.

Table 3 
Wasteful Software Development Expenditures Due to IPR Protection 

Clearly the projections in table 3 are little more than guesses.
More research would be needed to generate an accurate estimate of
the amount of software that is currently supported by IPR
protection. It should also be possible to find some basis for
providing better estimates of the amount of development that is
wasted on unnecessary duplicative activity as a result of IPR
protection. Assessing the relative rate at which software
development might proceed in an environment of free and open
software compared with IPR­protected software would almost
certainly require guesswork in any case, because we will never
see the two systems developing independently side by side.
Obviously, there is a loss due to secrecy, but there is no
obvious way of determining whether the numbers used in this table
capture the range of plausible estimates.

With these extremely important qualifications, the assumptions
used in constructing the table lead to scenarios in which between
28 percent and 72 percent of current software development
expenditures are wasted due to the distortions created by IPR
protection. The implication of such projections, if their basis
turns out to be plausible, is that a system of publicly financed
system of free and open software would require far less spending
on software development to accomplish the same results.
Alternatively, the same level of publicly provided spending could
lead to far more impressive accomplishments, if free and open
software were the standard.


Rent Seeking Activities of IPR Holders 

Another major source of waste associated with IPR protection in
software is the rent­seeking activities that it causes. The large
gap between the price and marginal cost of the product gives
producers far more incentive to advertise and market their
products than would be the case if they were selling at their
competitive­market price. While the additional sales that result
from such marketing may be highly profitable for the producers,
the expenses from sales efforts are largely a waste from the
standpoint of the economy as a whole. 4 

No reliable data exist on the amount of money devoted to
marketing relative to the amount spent on software development,
but it is likely that the ratio is quite high. In the
pharmaceutical industry, where patent protection leads to
mark­ups that average 300­400 percent above the free market
price, nearly twice as many people are employed in marketing than
in research. 5 The relative size of the monopoly profits due to
IPR protection in the software industry is even larger. Microsoft
reported spending roughly 30 percent more on marketing than it
did on research and development in the second half of 2004
(Microsoft 2005). If this pattern is repeated throughout the
industry, then the calculations used in the last section would
imply that approximately $15 billion a year is spent by the
software industry on advertising and marketing.

While excess spending on advertising and marketing is probably
the largest single rent­seeking cost associated with IPR
protection, it is just one of many. A second potentially costly
expense involves security measures that are necessary simply
because the software is subject to IPR protection. At the most
basic level, this involves designing digital locks, or other
features of the software, that make it difficult to copy. Some
portion of the resources devoted to software development must be
employed to making the software more difficult to duplicate,
instead of making it better for the consumer. Obviously, from the
standpoint of the economy, such expenditures are a complete
waste.

In addition to the resources devoted to designing software that
is more difficult to duplicate, software producers also must make
a wide variety of expenditures to ensure that their IPRs are
respected. These expenditures includes efforts to monitor the
distribution of their software over the web, legal fees
associated with enforcing IPRs, and contributions to political
campaigns and public relations efforts needed to sustain and
extend IPR protection.

Beyond the expenses incurred by producers, there are also costs
associated with enforcing IPR borne by other actors in the
economy. These costs include paying for the law enforcement
officials and the court personal required to enforce IPRs,
enforcement efforts imposed on third parties such as Internet
Service Providers and universities, and also restrictions on the
developments of technologies that could facilitate the violation
of IPRs.

This last category of costs is likely the most important. The
software industry (along with the entertainment industry) has
persuaded Congress to pass legislation that restricts the
production of hardware or software that facilitates infringement
on IPRs. 6 While the exact meaning of these restrictions has not
yet been clarified by the courts, clearly they have the effect of
slowing technological progress. If a hardware or software
producer must worry that the courts will prohibit them from
selling the product that they hope to develop, then there will be
less incentive to pursue the development of technology designed
to facilitate quick and accurate reproduction of digital
material.

The set of costs implied by these efforts to secure economic
rents are difficult to quantify. Software developers spend tens
of millions of dollars on lobbying and campaign contributions
every year. 7 They almost certainly spend an even larger amount
on lawyers engaged in work intended to either protect their own
IPRs or to defend the company against the charge of having
infringed on the IPRs of others.

There are no good measures of the cost that IPR holders (more
often in the entertainment sector than software) have imposed on
third parties by requiring monitoring or other efforts to
discourage IPR infringement. In some cases, that largest expense
may be simply be the time involved. For example, if a college
requires that 5,000 incoming students sit through a one­hour
lecture on respecting IPRs, the implicit cost would be $50,000,
if an hour of these students' time is worth an average of $10.

While it is virtually certain that search and copying software
and hardware would be more advanced if the development of these
technologies had not been impaired by efforts to enforce IPRs,
there is no obvious way to measure the extent to which
development of these technologies have been impeded. It is
reasonable to assume that we would see substantially more
progress in these areas if developers did not have to worry about
legal actions from holders of IPRs.

In sum, there are clearly substantial costs associated with the
rent­seeking of IPR holders in software. These costs are
extremely difficult to measure, but nonetheless are likely to be
substantial.


21st Century Alternatives to Patents and Copyrights 

As is the case with prescription drugs, the development of system
software and widely used and standardized applications is
probably best carried through by a central funding mechanism
controlled by experts in the field. 8 There are several key
principles that should be applied in setting up such a system.
First is the open source/ free software principle. The software
developed through this system is being paid for with the public's
tax dollars, this means that it belongs to the public. All
software produced through this system should be fully available
to the public to use and alter as they see fit (possibly subject
to "copyleft" restrictions -- anyone is free to use the software
directly or as basis for future modifications, as long as they
don't attempt to restrict its future use).

To enforce this principle of free software, it would probably be
necessary to require that anyone receiving funding through this
system be ineligible for IPR protection for any of their work for
a substantial period of time (e.g. 5 years). This would prevent
developers from gaming the system by being paid to develop
software, but then gaining IPR protection for their best ideas so
that they could personally profit from them. The great feature of
this system is that it would be largely self­enforcing. A
copyright or patent granted to a person who received support
through the Software Development Corps is simply invalid. Anyone
can freely ignore their IPRs. The sanction to the person who has
illegally claimed an IPR in such situations is simply that the
courts will not enforce their patents.

The second principle that it would be important to preserve is
competition. This can be accomplished by having the public
funding for software development divided among several competing
public corporations. For example, if the public funding for this
mechanism of software development is $2 billion a year, then this
money can be divided among ten firms that would each receive $200
million a year to support software development. These firms would
then be free to use this money to finance the software research
they deemed most promising, subject to the legal framework
described above. The public corporations could either look to
hire developers directly or subcontract out with firms or
research institutions, or some mix of the two mechanisms. (The
National Institutes of Health may provide a good model here. Less
than one­ fifth of its spending supports research that it
directly undertakes, the rest is contracted out with private
institutions and individuals.) 

At regular intervals (e.g. 7 years), the work of the software
development companies would be assessed by an independent
commission of experts and users. The worst two would be shutdown,
with two new ones being created to replace them. This should
ensure that the companies feel pressure to use their funds
effectively and that failed bureaucracies are not perpetuated
indefinitely.

In addition, it would also be desirable to have a separate pool
of money in this system (e.g. $100 million a year), which could
be used as prize money to reward outstanding breakthroughs in
software development. This should ensure that software developers
have substantial monetary incentives. Such a prize fund could
allow developers who do extraordinary work to get a bonus of
hundreds of thousands of dollars, or even millions of dollars, if
their work makes an exceptional contribution to the field.

Annual funding at the level of $2 billion a year ($2.1 billion
including the prize money) would be sufficient to hire 20,000
programmers a year, assuming base salaries of $80,000 a year and
25 percent overhead. 9 This sum would likely be sufficient to
replace the bulk of the software development currently supported
by the IPR system. Such a commitment of funding would be
relatively small for the federal government, being equal to
approximately 0.08 percent of annual spending. Furthermore, it is
likely to have most, if not all, of these expenditures reimbursed
in the form of lower­ priced computers and software. If
governments at all levels (federal, state and local) buy 5
million computers a year, and the average savings on new
computers and software came to $500 per computer, then the
savings to the government would be $2.5 billion a year,
considerably more than the sum committed by the federal
government to support software development. (Of course, much of
this savings would accrue to state and local governments.) 

In this case, the benefits to the rest of the economy, in the
form of much lower cost computers and software would be a free
lunch -- effectively the equivalent of a tax cut of between
$80­$120 billion a year, with no offsetting cuts in government
services. This would both lead to large immediate gains to
consumers and to a substantial stimulus to the economy. The money
that consumers save on buying computers and software can instead
be diverted to other areas of consumption or may be used as
savings to finance new investment.

IPR Alternative I: The Software Development Corps 

As is the case with prescription drugs, the development of system
software and widely used and standardized applications is
probably best carried through by a central funding mechanism
controlled by experts in the field. 10 There are several key
principles that should be applied in setting up such a system.
First is the open source/free software principle. The software
developed through this system is being paid for with the public's
tax dollars, this means that it belongs to the public. All
software produced through this system should be fully available
to the public to use and alter as they see fit (possibly subject
to "copyleft" restrictions -- anyone is free to use the software
directly or as basis for future modifications, as long as they
don't attempt to restrict its future use).

To enforce this principle of free software, it would probably be
necessary to require that anyone receiving funding through this
system be ineligible for IPR protection for any of their work for
a substantial period of time (e.g. 5 years). This would prevent
developers from gaming the system by being paid to develop
software, but then gaining IPR protection for their best ideas so
that they could personally profit from them. The great feature of
this system is that it would be largely self­enforcing. A
copyright or patent granted to a person who received support
through the Software Development Corps is simply invalid. Anyone
can freely ignore their IPRs. The sanction to the person who has
illegally claimed an IPR in such situations is simply that the
courts will not enforce their patents.

The second principle that it would be important to preserve is
competition. This can be accomplished by having the public
funding for software development divided among several competing
public corporations. For example, if the public funding for this
mechanism of software development is $2 billion a year, then this
money can be divided among ten firms that would each receive $200
million a year to support software development. These firms would
then be free to use this money to finance the software research
they deemed most promising, subject to the legal framework
described above. The public corporations could either look to
hire developers directly or subcontract out with firms or
research institutions, or some mix of the two mechanisms. (The
National Institutes of Health may provide a good model here. Less
than one­fifth of its spending supports research that it directly
undertakes, the rest is contracted out with private institutions
and individuals.) 

At regular intervals (e.g. 7 years), the work of the software
development companies would be assessed by an independent
commission of experts and users. The worst two would be shutdown,
with two new ones being created to replace them. This should
ensure that the companies feel pressure to use their funds
effectively and that failed bureaucracies are not perpetuated
indefinitely.

In addition, it would also be desirable to have a separate pool
of money in this system (e.g. $100 million a year), which could
be used as prize money to reward outstanding breakthroughs in
software development. This should ensure that software developers
have substantial monetary incentives. Such a prize fund could
allow developers who do extraordinary work to get a bonus of
hundreds of thousands of dollars, or even millions of dollars, if
their work makes an exceptional contribution to the field.

Annual funding at the level of $2 billion a year ($2.1 billion
including the prize money) would be sufficient to hire 20,000
programmers a year, assuming base salaries of $80,000 a year and
25 percent overhead. 11 This sum would likely be sufficient to
replace the bulk of the software development currently supported
by the IPR system. Such a commitment of funding would be
relatively small for the federal government, being equal to
approximately 0.08 percent of annual spending. Furthermore, it is
likely to have most, if not all, of these expenditures reimbursed
in the form of lower­priced computers and software. If
governments at all levels (federal, state and local) buy 5
million computers a year, and the average savings on new
computers and software came to $500 per computer, then the
savings to the government would be $2.5 billion a year,
considerably more than the sum committed by the federal
government to support software development. (Of course, much of
this savings would accrue to state and local governments.) 

In this case, the benefits to the rest of the economy, in the
form of much lower cost computers and software would be a free
lunch -- effectively the equivalent of a tax cut of between
$80­$120 billion a year, with no offsetting cuts in government
services. This would both lead to large immediate gains to
consumers and to a substantial stimulus to the economy. The money
that consumers save on buying computers and software can instead
be diverted to other areas of consumption or may be used as
savings to finance new investment.

The Artistic Freedom Voucher -- Leaving Software Support to the
Individual 

The second mechanism that could be used to support software
development is the Artistic Freedom Voucher (AFV), a system of
individual vouchers. 12 Under this system, every adult would be
provided with a certain amount of money (e.g. $100 dollars) which
they could use to support any person they like who is engaged in
creative or artistic work. Such work would include writing and
performing music, writing or performing in movies, writing books,
newspapers or designing software. The money could also be paid to
intermediaries who support creative or artistic work (for
example, an intermediary may specialize in supporting the writing
of mysteries or the production of a certain type of video game).

In order to be eligible to receive funding through the AFV
system, a person or intermediary would have to register with the
government in the same way that a non­profit organization or a
church must register with the government in order to obtain
non­profit status. In such cases, the registration amounts to
informing the government of what sort of creative/artistic work
the individual does or the organization supports. The government
does not assess the merit of the work, its only responsibility is
to ensure that fraud is not taking place, in the same way that it
verifies that religious organizations are in fact engaged in
religious activity and not running a business or a tax scam.

It may be desirable to have a separate software development
voucher, apart from a more general artistic freedom voucher, to
ensure that a certain portion of this funding go to software
development. The disadvantage of having this distinction is that
it would make the system more complicated. However, it would be a
possible modification if it turned out that individuals were
neglecting software developers in their allocation of funds from
an artistic freedom voucher.

As with the Software Development Corps, individuals who receive
funding through the AFV system (either directly or through an
intermediary) are prohibited from gaining IPR protection for
their work for a substantial period of time after receiving this
support. All the work must be placed in the public domain where
it is freely available.


The Competition Between IPR's and Alternative Mechanisms 

It is important to note that the IPR system can in principle
exist side by side in competition with these alternative
mechanisms for supporting software development. There is no
reason that software developers, like Microsoft, could not
continue to produce IPR­protected software and try to sell it in
the market. If they actually produce software that is
sufficiently superior to justify paying the IPR­ protected price,
then consumerswill still purchase it. These alternatives simply
provide a basis for a market test of the two systems.

In this context, it is important to understand that this
competition is between two formsofgovernmentsubsidies, not
between a government­funded system and a market­ based system. In
one system, the government pays for the
developmentofsoftwareupfronteither through theSoftware
DevelopmentCorpsor theAFVsystem. In the IPR system, the
governmentgives software developers a state­imposed
monopoly,where the governmentcriminalizescompetition for a period
of time.

In a free­market system, software developers would simply be left
on their own -- producing their software and selling it as best
they could. If another software developer had the ability to
reproduce and sell the software, then the original developer
wouldhavenomechanism to prevent such sales. IPRprotection is
oneway that the government interferes in the free market, in this
case, in order to provide incentives to
softwaredevelopers.Theappropriatequestion is whether IPRs are the
best mechanism.


Conclusion 

Relying on IPRs as the major means for supporting the development
of software brings with a large cost in terms of economic
inefficiency. The most obvious source of inefficiency is the
enormous mark­up over the free­ market prices that results from
the government enforcement of monopolies in the sale of computer
software. For the most part, software that currently commands
substantial prices as a result of IPR protection would be
available at zero cost over the Internet, in the absence of such
protection.

However, IPR protection also distorts the process of software
development itself. It encourages unnecessary duplication, as
software developers devote time and resources to duplicating
software that is protected by IPRs in order to gain a portion of
the monopoly rents. In the absence of IPR protection, effort
would generally be better directed towards improving existing
software. Secrecy and restrictions on the use of IPR­protected
software impede the process of software development as well.

IPR protection also leads to large expenditures by IPR holders to
protect their rents. These expenditures include advertising and
marketing, legal fees, and lobbying and campaign contributions.
All ways in which IPR holders seek to maximize their monopoly
rents. IPR holders also impose costs on others, such as Internet
Service Providers and universities, by requiring them to take
steps to enforce IPRs. In addition, IPRs have managed to obstruct
the development of software and hardware that facilitates
searches and reproduction of digital material.

The paper proposes two alternative mechanisms for funding
software development, both of which can co­ exist in competition
with the IPR system. One would provide approximately $2 billion a
year of government funding through a set of competing software
development companies. In this case, all the software would be
made freely available to be reproduced or altered, subject to
copyleft principles. The second mechanism would allow all adults
a certain sum of money (e.g. $100) which could be used to finance
any creative or artistic venture of their choice, including the
development of software or computer games. Any work supported
through this system would also be immediately placed in the
public domain.

It is quite likely that the government could fully recoup the
money it paid for software development through lower prices on
the computers and software it purchases. In this case, the lower
computer and software prices seen by the general public would
have the same effect as a tax break in the neighborhood of
$80­$120 billion a year, with no offsetting cut in public
services. Money that had been spent paying for computers and
software would instead be devoted to either items, providing a
substantial boost to the economy and leading to millions of new
jobs.


Appendix 

The calculations in Table 1 assume that the world sales for
desktop computers were 80 million units in 2004. They assume that
sales of laptops were 60 million. The average prices are assumed
to be $800 and $1,000, respectively. It is assumed that without
IPR protection, the average price of the computers purchased
would fall by $200, and that the value of the additional software
included with an average computer be $600 at current prices, with
a value to consumers equal to $300. (In other words, it is
assumed that the decline in the price of computers is equivalent
to an average of $500 per computer.) It is assumed that the
average price of a computer or video game is $29.50, based on
data from the industry. 13 The low, medium, and high elasticity
scenarios assume a constant elasticity of substitution utility
function with elasticities of 0.2, 0.45, and 0.7, respectively.
The estimates of current sales volumes are loosely derived from a
recent article in the business press which put worldwide sales of
notebook computers at 60 million units a year. 14 It is assumed
that desktop sales are somewhat larger, at 100 million units a
year. The estimate of the volume of computer and video games
sales at 250 million units a year at an average price of $29.50
is taken from Essential Facts About the Computer and Video Game
Industry, 2005, Entertainment Software Association
[www.theesa.com/files/2005essentialfacts.pdf]. The calculations
in Table 2 present sums from Table 1.

---

Footnotes:

1 There is an important issue about how best to keep software
developed with public funds in the public domain in a world where
private individuals can still get IPRs. One possible mechanism is
the "copyleft" system developed by the Free Software Movement.
Under the copyleft system, the software is copyrighted, but can
still be freely reproduced, as long as an subsequent refinements
of the software are made freely available either by being placed
in the public domain or being subject to a new copyleft. A fuller
description can be found at http://www.gnu.org/copyleft/gpl.html 

2 This estimate is derived from the Food and Drug
Administration's classification of new drugs and the
pharmaceutical industry's estimate of the cost of developing
duplicative, as opposed to breakthrough drugs, see "Financing
Drug Research: What Are the Issues?" 

3 Microsoft's annual statement from the second half of 2004 can
be found at [http://www.microsoft.com/msft/earnings/FY05/
earn_rel_q2_05.mspx].

4 Advertising and marketing can provide information about
products, which is economically beneficial. However, if this
information is misleading, and prompts consumers to opt for
software that is less well suited for their purposes, the sales
effort contributes negatively to the economy. Even if no
resources were required for the sales effort, the economy would
be better off without a misleading sales effort.

5 (http://www.pharma.org/publications/publications/profile01/
app_a3.phtml) 

6 This discussion refers to the Digital Millennium Copyright Act,
which was signed into law by President Clinton in 1998.

7 The Center for Responsive Politics reports that Microsoft along
spent more than $6 million on lobbying in 2004 (www.crp.org).

8 This model is based on a proposal for financing prescription
drugs, which was part of a bill recently introduced in Congress
(The Free Market Drug Act of 2004). A description of this
mechanism can be found in "Financing Prescription Drug Research:
What Are the Options?" 

9 The Bureau of Labor Statistics Occupational Employment
Statistics Survey estimated the average annual wage of a computer
programmer in 2004 at approximately $80,000
[http://bls.gov/news.release/ ocwage.t01.htm].

10 This model is based on a proposal for financing prescription
drugs, which was part of a bill recently introduced in Congress
(The Free Market Drug Act of 2004). A description of this
mechanism can be found in "Financing Prescription Drug Research:
What Are the Options?" 

11 The Bureau of Labor Statistics Occupational Employment
Statistics Survey estimated the average annual wage of a computer
programmer in 2004 at approximately $80,000 [http://bls.gov/
news.release/ocwage.t01.htm].

12 This mechanism extends a proposal for supporting creative and
artistic work described in "The Artistic Freedom Voucher:
Internet Age Alternative to Copyrights" 

13 http://www.theesa.com/files/2005essential facts.pdf 

14 These calculations assume 2005 world sales of notebook
computers are 60 million, at an average price of $1000 per
computer. This is derived loosely from recent public sales data
(" Taiwan took 81.3% share of global notebook PC sales in Q2"
[http:/
/english.www.gov.tw/index.jsp?id=12&recid=109038&viewdate=0 ]).
The calculations assume that the 2005 sales volume of desktop
computers is 80 million at an average price of $800.