Policy Pitfalls: The Use and Misuse of R&D Indicators
It is worth saying something about the pitfalls of R&D as a policy indicator,
especially via the most widely-used indicator, that of "R&D Intensity." This is the
ratio ofR&D expenditure to some measure of output. For a firm, it is usually the
R&D/Sales ratio. For an industry or a country it is the ratio of business expenditure
on R&D(often known asBERD)to total production or value added. For a country it
is usually gross expenditure on R&D (GERD) to GDP.
The R&D/GDP ratio is used in two primary ways. First, it is used to characterize
industries-high BERD/GDP ratios for an industry are held to identify hightechnology
activities. Second, a high GERD/GDP ratio for a country is often believed
to indicate technological progressiveness and commitment to knowledge creation
(seeGodin 2004 for an account ofthe historical background to these notions).
For countries, there is a distribution of GERD/GDP intensities, as Table 6.1
indicates. Both analysts and policy makers often treat a particular place in the.
ranking, or the GECD average, or some particular GERD/GDP ratio as desirable
in it self.So Canada, for example, has the objective ofraising its ranking to fifth in the
GECD table; Norway has the target of reaching the GECD average for GERD/GDP;
and the ED as a whole has a target of reaching a GERD/GDP ratio of 3 per cent (it
could be argued that this target dominates ED technology policy making at the
present time). But what is the indicator really telling us?
A basic problem is that R&D intensity depends on the industrial mix. Currently
the GECD uses a four-tier model to'classify industries, in which the basic criterion is
the BERD/Production ratio:
high-tech industries
medium high-tech industries
low-tech industries
low-tech industries
>5%
>3%
>1%
>5%
>3%
>1%
>0%
R&D/Production
R&D/Production
R&D/ Production
R&D/ Production
Since industries vary considerably in their BERD/GDPratios, the aggregate BERD/
GDP ratio may simply be an effectofthat fact that industrial structures are different
across countries. Acountry or region with large high-R&D industries will naturally
have a higher aggregate BERD/GDPratio than one with most ofits activities in lowR&
D industries. These structural issues largely explain the differences in R&D
intensities across large and smaller economies (Sandven and Smith 1997). The
question then is, does a specific industrial structure really matter? This is question
for debate, which cannot be addressed let alone settled here (it is interestingly
explored in Pol et aL 2002); however the desirability of specific industrial structures
is the real issue underlying use of this aggregate indicator, though it is rarely
explicitly discussed. It is worth noting also that within an industry there tends to
be a wide distribution of R&D intensities among firms, so it is common to find