FLYING SOFTWARE: Is the Information Society heading
South?

In 1998, developing countries will export around US$3bn-
worth of computer software to Western markets, making use of
telecommunication networks: a major signal that Information
Society benefits can be global.  However, such 'headline
images' are deceptive.  Software production retains output,
location and skill skews that provide limited benefits for DCs.
These skews will be difficult to eliminate, but government
action is needed to address them.

Over the past five years, all DC software exporters worth their
salt have invested in telecommunication links and many are
Internet-connected.  The new image projected is one of 'virtual
development' in which clients sitting in the West interact with
software professionals developing packages overseas.  The
reality, though, is somewhat different with most software
export contracts involving:
>>A skewed output profile: software packages represent a
tiny proportion of total DC exports.  Thus, the Microsoft of the
21st century is not currently being born in the back streets of
Bangalore.  Instead producers in the South provide software
services.
>>A skewed location profile: large amounts of
development work take place at the client's site, i.e. by having
DC software developers fly over to work with the client.
>>A skewed skills profile: most work undertaken by DC
developers is relatively low-skill software construction and
testing, leaving the high-skill tasks of analysis and design
residing in Western hands.

As a result, much of the US$3bn earned leaves developing
countries to pay for: travel and living allowances of the DC
developers who work at the client site; marketing expenses;
information and communication technology imports used for
DC-based contract components; and profit repatriation by the
many multinationals involved in this trade.

The dominance of service over package exports can be
explained by the almost insuperable barriers of cost, skills, and
information that exist in the package sector.  But why do the
other skews persist?  Three factors can be highlighted:
>>The 'Information Superhighway' is not a reality for most
DC users.  Fax and email correspondence, not
videoconferencing, remain the backbone of communications.
These mechanisms cannot provide the depth of interaction that
software development requires.
>>Client uncertainty about DC firms' skills, capability and
credibility.  To reduce risk, clients choose to retain as much
control as they can over production.  This means only
contracting out relatively simple, low-skill tasks, and having
work carried out under their noses.
>>A 'programmer heavy' skills profile in developing
countries, with no shortage of raw recruits but very few
experienced project managers, analysts and designers.

How can developing countries break out of this 'body shopping
prison'?  The spread of global networks will help, but only
addresses the first factor, not the second two.  Time will also
help, as individual client-developer relationships move slowly
up a 'trust curve' and uncertainties diminish.  But, finally, there
is a role for the state.

The Department of Defense was a key driver in the creation of
the US software industry, and all subsequent development in
other countries has been state-initiated, state-led or state-
promoted.  Selective policy liberalisations - such as removal of
software import tariffs - have a role.  However, the elimination
of all state interventions and the free play of market forces will
lead only to the atrophy of local software-related technological
capabilities.

From South Africa to Egypt to India to Singapore the story is
the same: building a viable software industry requires
government promotional interventions in financing and
marketing, in skills and infrastructure development, in
procurement, and in the diffusion of best practice.

REFERENCES:
Heeks, R. (1996) India's Software Industry, Sage Publications
Heeks, R. (forthcoming) Building Software Industries in
Africa, IDPM (University of Manchester) Working Paper

---------------------------------------------------

Dr Richard Heeks, Institute for Development Policy and
Management, Precinct Centre, University of Manchester,
Manchester, M13 9GH, UK
T: +44-161-275-2870
F: +44-161-273-8829
E: [log in to unmask]
W: http://www.man.ac.uk/idpm/