IEEE Technology and Society Magazine - March 2023 - 38

Feature
has its limitations, compared to seven base units in
the International System of Units (SI) and a multitude
of secondary units for natural sciences.
A relatively new and fairly narrow research literature
indicates that many innovative improvements
in performance/functionality are not only missed by
measured GDP but may, in fact, sometimes lower
it. While this contention is still controversial among
many economists, an influential overview by Feldstein
[2] concluded that measurement problems,
even within the framework of what GDP intends to
measure, have grown to the point where the GDP
measure can, at best, be seen as a minimum measure,
with a potentially very high degree of underestimation.
In particular, he argued, the importance of
innovation was underestimated. We would add that
underestimation may have been greatly exacerbated
by the pandemic.
Since then, a number of studies, elaborated further
below, have also been published that further
indicate that significant value creation takes place
beyond what national accounts intend to capture
through the GDP measure in the form of consumer
surplus and various externalities.
The pandemic has accelerated the kind of digital
work that falls under the definition of household
production which has never been included
in GDP, and the kind of digital consumption with a
zero price tag which is not counted either. Perhaps,
more importantly, the pandemic has underlined the
importance of health-related innovation whose benefits
also are rarely registered in GDP. Similarly, the
value generated by climate-oriented innovations is
largely excluded. More generally, efforts to prevent
or prepare for calamities are not well captured in
GDP or other traditional measures of the returns to
innovation.
An Einsteinian " Gedankenexperiment, " thought
experiment, can illustrate this. Assume that a
research group managed to invent a vaccine against
all
Gap between the performance
development of digital systems and
registered value added
The most well-known quantified model that
describes technological performance enhancement
is Moore's law, which states that the number of transistors
in computer chips has doubled every second
year, followed by the corresponding computing performance
increase (Figure 1).
The rapid development of electronics is behind
the qualitative advance in many other areas.1 For
example, the price of a cell phone in the early 1990s
was equal to or higher than today, while performance
and functionality are not comparable. The
processor in an iPhone today has higher computing
performance than a Cray 2, which was the fastest
supercomputer when digital cellular telephony was
introduced at the beginning of the 1990s. At the time,
a Cray 2 cost approximately $20-30 million and consumed
200 kW of electric power.
Other areas of technology show similar trends
and have together contributed to the rapid pace of
digitalization, such as the development of optical
and photonic technologies resulting in fiber-optic
cables that are the backbone of almost all communication
networks (Figure 2).
However, the exponential performance growth
diseases, including cancer and autoimmune
afflictions. The result is that more than half the medical
sector is wiped out. Close to 10% of GDP would
disappear, at least until unemployed doctors, nurses,
and so on find something else to do. Should this
really be accounted for as an economic disaster as
national accounts would do today since they largely
ignore the quality of healthcare improvements... or
a major improvement for humankind?
38
of the technologies, fundamental to digitalization,
has not been reflected in the revenue of the leading
companies. Examples include Intel, a leading company
in the computer chip industry, and Ericsson,
a leader in the mobile telephony infrastructure. The
reason to choose these enterprises is their leading
role during the relevant time period in the respective
technology field. The performance of their products
increased 105-106 times since 1990. During that
time, their markets have grown significantly and
have been characterized by some oligopoly, with
few competitors. Nevertheless, the revenue of these
companies increased only few times between 1990
and today, and the profits are rather normal in the
business world. This example shows that the performance
increase of products and services by many
orders of magnitude does not translate into the monetary
value captured by the GDP measure. Yet, without
this performance, new innovative products such
as autonomous vehicles would be impossible. The
DRIVE Thor chip from NVIDIA for next-generation
1Swanson [4] gives a good description of the technical gains in " Moore's law at 50. "
IEEE Technology and Society Magazine
IEEE Technology and Society Magazine - March 2023

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - March 2023
