How microeconomically insignificant is Defense R&D?

[world!srctran@uunet.uu.net (Gregory Aharonian)]

 
     For some time I have been arguing that many DoD software initiatives
are a waste of money and should be no longer funded, for the following
reasons - they seem to have little impact outside the DoD (STARS), they
waste money on extravagant services (ASSET), or they duplicate services
the private sector can more cheaply offer (SEI).

     But for all of my arguments, especially in light of the DoD's attempts
to start dual-using everything, the real question is:  How microeconomically
significant is DoD software research?  What exactly is this country getting
for the hundreds of millions of dollars being spent each year?  And more
generally, how microeconomically significant is DoD research?
 
     To shed some light on this, and to give ammunition to those of us who
believe that DoD, and the rest of the government, should not be in the venture
capital business, (i.e. stop using my tax dollars to compete with me - that's
what we fought the Cold War for) consider the abstract to the following
article:
 
 
		DEFENSE R&D, TECHNOLOGY AND ECONOMIC PERFORMANCE
		  A LONGITUDINAL ANALYSIS OF THE US EXPERIENCE
		     by Alok Chakrabarti, C. Lenard Anyanwu
		       New Jersey Institute of Technology
            IEEE TRANS ENGINEER MANAGEMENT 40, 2, 136-145, May 1993
 
    This paper examines the issue of impact of defense expenditure from
    different perspective, i.e., in terms of the direct relationships
    between defense R&D and economic performance as well as the indirect
    relationships via the development of (1) technical and scientific
    skills and (2) new technology.  The model was estimated for the
    period 1955-1988 on a time-series set measured as elasticities. This
    effect of defense R&D is observed particularly through technological
    change as measured by the number of patents granted to US organizations
    and individuals.
 
    There is no statistically significant evidence of resource diversion
    or "crowding effect" on the civilian economy due to defense R&D.
    SIMILARLY, THERE DOES NOT SEEM TO BE ANY STATISTICALLY VISIBLE
    EVIDENCE OF DIRECT EFFECT FROM DEFENSE R&D TO THE ECONOMY. Interestingly
    the non-R&D aspect of defense spending appears to have no statistically
    significant effect on the major components of civilian economic
    performance, technical-skills formation or technological change.  From
    a policy point of view, this suggests that technical spillovers may
    be limited to a specific kind of defense spending and not to defense
    spending in general.  Another interesting implication is the rivalry
    between R&D and non-R&D defense spending is in favor of the latter.
 
===========================================================================
 
    Thus one of the few independent studies of defense R&D spending, unlike
those SDI (and others in the DoD) commission to brag about million dollar
spinoffs from billion dollar investments (can you spell ROI?), we see
evidence that maybe the country is getting little in non-defense return
from the billions of taxdollars being spent each year.

    For software R&D, well VHDL, SGML, CALS, TCP/IP came out of the DoD
and have had much benefit on their communities.  Ada, probably the most
spinoffable, is dead outside the DoD, while much DoD software engineering
methodology (other than maybe SEI's CMM) that has been developed over the
last fifteen years rarely was adopted by the software industry.  DoD software
reuse efforts are economically unsound and function only as long as the
DoD provides welfare money - none know how to transition to private operation,
so not much spinning off there.  Most government software that has been
commercialized has come out of NASA and DOE.  And the music world has not
benefitted much from DoD's subsidy of the Ada Follies.  And ARPA, well since
ARPA doesn't have a directory of all of the software it has funded, it
certainly can't determine how much impact their software has had.  So an
honest study of DoD software R&D's impact on the general economy would
probably reflect the conclusions of the above paper.
 
    In the case of defense R&D for software, given that such software
rarely if ever: appears or transitions to the private sector, is patented,
is used, or is commercialized, it becomes interesting to start speculating
on whether the DoD should stop funding its software R&D activities, and
rely more on the private sector, which for the most part is years ahead of
DoD efforts anyways, and without the spending of taxpayers dollars needed.

    Anyways I highly recommend people get a copy of the article.

-- 
**************************************************************************
 Greg Aharonian                                      srctran@world.std.com
 Source Translation & Optimization                            617-489-3727
 P.O. Box 404, Belmont, MA 02178