3. TECHNICAL CHANGE
3.1 How technologies evolve
Schumpeter (1942) distinguishes three stages in the process by which a new, superior technology
permeates the marketplace. Invention constitutes the first development of a scientifically or technically
new product or process. Innovation is accomplished when a new product or process is made available on
the market. Diffusion (or dissemination) is the process that sees a successful innovation gradually coming
to be widely available for use in relevant applications through adoption by firms or individuals. The
cumulative economic and environmental impacts of new technology results from all three of these stages,
which we refer to collectively as the process of technological change. (Jaffe et al., 2002).
However, these definitions might wrongly suggest that technical change is a linear process that simply
goes from invention to innovation to diffusion. In fact, it is more a cyclical process; the feedback between
market experience and further technical development are especially important. Market prospects are the
most vital stimulant of industry research and development (R&D) and the deployment of technologies is a
key source of information on them. Market development and technology development go hand in hand
(IEA, 2003).
The view that technology deployment in the marketplace – not only research and development efforts –
is a
key element to speed up technical change, is borne out by lessons from past technological developments.
They reveal that the costs of technologies decrease as total unit volume rises. A metric for such change is
the “progress ratio”, defined as the reduction of cost as a consequence of the doubling of cumulative
installed technology. This ratio has proven roughly constant over time for individual technologies –
although there is considerable variance in the ratio between technologies. The fact that the progress ratio is
usually constant means that technologies learn faster from market experiences when they are new than
when they are mature. The same absolute increase in cumulative production has a more dramatic effect at
the beginning of a technology’s deployment than it has later (see IEA, 2000). This is why new techniques,
although more costly at the outset, may become cost-effective over time if they benefit from sufficient
dissemination. Figure 1 below shows this phenomenon in the power-generating sector.
COM/ENV/EPOC/IEA/SLT(2003)4
13
Figure 1. Electric Technologies in EU 1980-1995
0.01
0.1
1
10
0.01 0.1 1 10 100 1000
Cumulative Electricity Production (TWh)
Cost of Electricity (ECU(1990)/kWh)
Photovoltaics (~65%)
Electricity from
Biomass (~85%)
Supercritical Coal (97%)
NGCC (96%)
Wind Power - Average (82%)
Wind Power - Best Performance
(82%)
1985
1995
1980
1995
Numbers in parenthesis are estimates of progress ratios. They indicate the change in cost when market size doubles.
Thus, for example, if the size of PV markets double, the cost of PV electricity is reduced to 65% of its previous value.
Source: Source: IEA, 2000
Projecting forward using concepts of learning-by-doing is risky; there is a clear indication that some
mature technologies no longer follow the progress ratio of their early development. However, the concept
may well be robust for many of the new, immature technologies such as renewable energy. Using these
ratios, it is possible to gain some idea of how competitive advantages may change with time. For example,
in photovoltaics, a break-even point with fossil fuels might be expected around 2025 if historical growth of
PV deployment at 15 per cent per year continues. The difference in price between these technologies and
least cost options constitute “learning investments” – cumulative costs for supporting new technology. The
area under each curve (or line) can be used as a rough estimate of the total capital investment required to
bring the technology’s cost down to a certain level.
Much of the progress in PV growth is supported through niche markets – in remote places where
photovoltaics are already the most cost-effective solution. In the case of wind power, market deployment
has increased in a number of countries where policies drive consumers to pay extra for wind power. The
US in the 1980s, and Denmark in the 1990s were the main leaders, and more recently Germany, Spain and
India have seen extensive growth in wind generation. Whether these technologies can continue to have
rapid increases in cumulative capacity will in part depend on their moving from small niche markets to
more mainstream application.
Distinguishing the effects of R&D efforts and those arising from market deployment may not be easy, as
Clarke & Weyant (2003) point out. Learning curves literature usually misses a detailed history of R&D
expenses, while R&D literature often ignores learning effects. Moreover, the coexistence of increased
market shares and decreased costs does not necessarily demonstrate that the former caused the latter. The
causality relationship works both ways: when cost decrease, niche markets increase.
COM/ENV/EPOC/IEA/SLT(2003)4
14
3.2 Path dependence
This feedback process from markets to technical improvements, creating increasing return, has numerous
consequences. It tends to create “lock-in” and “lock-out” phenomenon: it is not because a particular
technology is efficient that it is adopted, but rather because it is adopted that it will become efficient
(Arthur, 1989). Technological paths often depend on initial conditions. Thus, technologies having a small
short-term advantages may “lock-in” the technical basis of a society into technological choices that may
have lower long-term advantages than others technologies, which are consequently “locked-out”.
The systemic and cumulative nature of technological change leads to clustering effects, or technological
interdependence, and possible phenomena of increasing returns: the more a technology is applied the more
it improves and widens its market potential. Change is directionally persistent based on an accumulation of
past decisions. As noted by Roehrl and Riahi (2000), “technological change can go in multiple directions,
but once change is initiated in a particular direction, it becomes increasingly difficult to change its
course.”
However, trends are neither immutable nor infinite: increasing returns might be bounded and scale
economies exhausted. Entrepreneurs facing new pressures may break technology barriers, and technologies
that lost out in early competition may subsequently become successful. For example, while electric cars
seem to have lost the battle against internal combustion engines a long time ago, they might be given a
second chance –
although their competitors now benefit from about a century of technical improvements.
Several policy conclusions may be drawn from this analysis. The first is that redirecting technical change
in order to reduce CO2 emissions has to build on “learning-by-doing” processes. Laboratory research and
development efforts are unlikely to be sufficient to produce sufficient progress so as to impose new
technologies on competitive markets in one shot. As detailed below, creating markets for new energy
technologies necessitates broader efforts.
Another, no less important policy conclusion, has to do with the timing issue. As pointed out by Roehrl and
Riahi (2000), “research, development and demonstration efforts as well as investment decisions in the
energy sector over the next two to three decades are critical in determining which long-term technological
options in the energy sector may be opened, or which ones may be foreclosed”.
3.3 Technological and behavioural change
Another dimension of the debate on technology in mitigating climate change relates to the important role
that behavioural change can play in reducing emissions. For some, technical change should simply
suppress emissions, and nothing else should change. For others, there is almost no need for new
technologies, as changes in behaviour could accomplish large emission reductions. These extreme and
opposite views are further detailed below.
3.3.1 Two polar views
The role of technology versus that of behaviour may be characterised (and caricatured) in two polar views.
According to the first view, technology should solve the problem by itself. Its role would be to provide, in
particular, energy systems with no or very low carbon emissions. While economic development would
proceed in a sustainable manner, no behavioural change would be needed or implied to get this result.
COM/ENV/EPOC/IEA/SLT(2003)4
15
Perhaps representative of this view is the recent paper by a group of US scientists reporting in Science
(Hoffert et al., 2002): “Primary power consumption today is ~ 12 TW, of which 85 per cent is fossilfuelled. Stabilisation at 550, 450 and 350 ppm CO2 by Wigley et al. Scenarios require emission-free power
by mid-century of 15, 25 and >30 TW, respectively. (…) Arguably, the most effective way to reduce CO2
emissions with economic growth and equity is to develop revolutionary changes in the technology of
energy production, distribution, storage and conversion.”
The opposite view would give a less important – or even trivial – role to technical change. In this
worldview, changes in behaviour would achieve a significant part of the emission reductions. People
would reduce their use of carbon-intensive materials and services. They may, for example, travel less for
leisure and spend leisure time and money on different and less carbon-intensive activities.
A good representative of this second vision is Jancovici (2002), although he does not dismiss a role for
technology. He sees sustainable and carbon-free energy being mainly derived from nuclear and solar power
(and its derivatives); he estimates that together (over the long term) these would provide about half of
today’s primary energy. As a consequence, in order to meet CO2 targets, per capita energy consumption in
industrialised countries would need to be significantly reduced. To stabilise CO2 concentrations below 450
ppm implies that many commodities or services would be available in lesser quantities than today – in
particular such highly energy-intensive commodities as mass air and road travel, inexpensive concretebased construction for housing or out-of-season fruits.
This argument echoes that made by developing countries experts or negotiators that “luxury” consumption
should not be given the same weight in analyses as people’s “basic needs”. It also relates to the argument
of “philosophers” as noted by Grubb (1997): “We can and should learn to become less obsessed with
material consumption and the burning of carbon that goes with it, and that by putting more emphasis upon
the quality of our physical and social environment we would anyway be better for it”.
An example of these two trends may help clarify the differences. Using transport, and according to the
first vision, fuels would become carbon free. They would be manufactured from nuclear or renewable
energy, or carbon emitted would be captured and stored at some point in the refining process. This may
require deep changes in the way cars and trucks are designed and built – such as replacing current motor
technology with fuel cells – and would allow people to keep their travel and commuting habits unmodified.
According to the second vision, new standards for energy efficiency should be imposed on carmakers,
which could modify the size and the weight of most cars. Moreover, use of cars should be regulated, either
by higher fuel prices or by restrictions for use in cities, or both; more efficient mass transit systems should
be given priority in land-use and public investments, so as to encourage people to switch transport modes.
Even more, people should be induced to reduce travel – both in terms of distance and frequency, through
pricing or other policies. Urban planning and related policies (such as, for example, credit policies for
construction work) would tend to maintain or increase density in urbanised arenas, not expand them.
