From trade theory it follows that as a result of comparative advantages countries specialise in certain
sectors. As a consequence, differences in productivity growth between countries and sectors are also
likely to occur. Through international trade and relative price changes all countries profit from this
specialisation pattern. In competitive domestic and international markets, productivity growth in a sector
will be transferred into lower output and export prices. For example, the Netherlands contains a large
financial and services sector instead of an ICT-sector, but the enormous productivity gains in the ICT
sector have lead to welfare gains in the Netherlands through lower prices of computers.
The importance of terms of trade effects caused by (productivity) growth is empirically quite significant.
For instance Acemoglu and Ventura (2001) find that a 1%-point faster growth is associated with a 0.6%-
point deterioration in the terms of trade. Nahuis and Geurts (2004) find significant effects of productivity
growth on relative export prices of industrial goods. Within a year, about a quarter of productivity growth
in a country is translated into lower export prices. In a four year period, this share has increased to almost
60%, whereas over a 25 year period 90% of productivity growth has been translated into lower export
prices.
Bayoumi and Haacker (2002) investigate the welfare impact of the recent productivity surge in the ICT-
sector for 29 countries. They find that welfare benefits mainly accrue to users of ICT, not to producers,
because of falling relative prices. Of the five most specialised ICT countries only Singapore has
significant welfare gains. The other four countries with a yearly contribution of ICT of more than 1%
point to real GDP growth (Ireland, the Philippines, Malaysia and Thailand) suffered such terms of trade
losses that the growth of welfare is lower than 0.20% a year. On the other hand, some of the non-ICT
producing countries, such as Denmark and Australia, have positive contributions to welfare growth due
to an improvement in their terms of trade.
Of course, comparative advantages and the sectors that countries specialise in are subject to change.
However, an explicit strategy to promote the development of certain sectors is not without risk. Not only
does the government experience information problems about the optimal specialisation pattern, but if too
many countries try to specialize in the same sectors with high productivity growth, it may also result in
excess supply and inefficient allocation.
To summarize, the welfare effects of a relatively high productivity growth are smaller than expected at
first sight. In the long run, the welfare effects of productivity growth are partly offset by a deterioration
of the terms of trade of a country.
3. Market and government failures in the market for innovation
Legitimization of innovation policy starts by identifying market failures that may lead to suboptimal
outcomes for the society as a whole, because external effects may drive a wedge between private and
social returns. Theoretical and empirical literature point to both negative and positive external effects of
R&D (Jacobs et al, 2001). Firstly, insufficient market power may limit the ability of a firm to internalize
all benefits of innovation or knowledge accumulation (rent spillovers). Furthermore, the mobility of
researchers and the inability to keep innovations in new products a secret for competitors create
knowledge spillovers (Jaffe, 1996). Also, negative spillovers might occur because an innovation may
reduce the profits of competitors. On average, empirical evidence points to positive spillovers.
Government policy aimed to reduce the gap between private and social returns could increase welfare. In
addition to external effects, imperfect information, network effects and market power of incumbent firms
may justify policy intervention as well. These market failures may decelerate the diffusion of knowledge,
although this is (partly) offset by the monopoly rents of early adopters of new technologies (CPB,
2002a). On the basis of these market failures, there may be room for government intervention.
However, the effectiveness of innovation policy depends on the occurrence of government failure.
Limited information, limited control over private responses and limited control over bureaucracy can
reduce the effectiveness of policy intervention. The risk of ineffective decisions is increased by lobby