| 1 |
The 1993 SNA also recommended treating large databases as capital
assets. Database software either purchased (e.g. Oracle, Microsoft
SQL Server) or developed on own-account have been capitalized here,
but the database content, its creation and its updating have not.
This latter remains for future consideration.
|
| 2 |
Software refers in general to the encoded instructions executed
by electronic devices including computers for performing operations
and functions. This includes both systems software and user tools
(operating systems, network control, performance measurement and
job accounting tools, utilities, compilers, CASE tools, etc.) and
applications software (CAD/CAM, reservation management systems,
word processing, spreadsheets, payroll systems, etc.).
|
| 3 |
Making the distinction between these three types of software
is easier said than done. Software developed ‘in-house’
for own-use may have viable applications elsewhere and may be sold,
leased or licensed to other organizations, blurring the distinction
between own-account and custom software. Specialized software may
also be integrated with more generalized software packages, blurring
the boundary between custom and pre-packaged software.
|
| 4 |
Statistics Canada’s industry surveys capture expensed software
under a catch-all category for ‘other office supplies’,
with the Survey of Computer Services (SCS) being one of the few
to explicitly provide for software.
|
| 5 |
It can be difficult to make these distinctions and remove such amounts
in practice. In the case of own-account repair and maintenance of
software, related costs are deemed to be removed through an adjustment
for the time computer programmers and systems analysts spend on non-software-development
tasks. With respect to purchased software repair and maintenance,
training, systems and technical consulting and other professional
services, these are not removed when they are priced in with, and
an integral part of, a software sales contract. When such services
are purchased separately from software they are excluded. Repairs
and maintenance, however, is an exception. The SCS gathers information
on industry revenues from a number of specific services, but revenue
from software repairs and maintenance is not explicitly asked for,
nor is any guidance given to respondents on where to report it. It
may be reported as revenue from custom software development, but it
could also be reported under a catch-all category for other professional
services. To the extent it gets reported as the former, some repair
and maintenance gets included as investment in custom software. The
1999 SCS has introduced a new revenue category for after-sales support
and maintenance which may shed some light on this issue. |
| 6 |
This includes license fees paid for the use of software in production,
regardless of the term of the license agreement (although, almost
all are thought to be for terms of one year or more). A de facto transfer
of ownership rights is assumed here, and the licensee is deemed to
make an investment in software despite the fact there is no legal
transfer of ownership. This treatment better reflects the economic
reality and helps in economic analyses, for instance, of productivity
by industry. Still, the question has arisen as to whether license
fees should continue to be treated as intermediate in the case of
software, with the licensor, not the licensee, viewed as having made
the investment. A joint OECD/Eurostat Task Force has been set up to
consider this issue among others related to software investment. See
OECD, “Software Measurement: Issues Paper,” OECD Meeting
of National Accounts Experts, Paris, September 2001. |
| 7 |
More precisely, starting in 1988, respondents to CAPEX were asked
to include capital spending on software along with that on hardware
under a combined hardware and software asset category that has been
associated typically only with ‘hardware’ and treated
as such. In the first year software was explicitly identified (CAPEX
1998), roughly $2.8 billion was reported as capital spending on software,
representing almost one-quarter of reported capital spending on both
software and hardware combined. Based on the estimates here, business
and government software expensed was three times the amount of capital
spending on software in that year. |
