Table 1, is further shown in Figure 2B. Coupling Figures 2A and 2B, we see a remarkable
consistency across these groups of innovators, suggesting a precise and general phenomenon: a
sharp decline in early life-cycle innovative output.
A natural mechanism for declining innovative output in the early life-cycle is a
corresponding increase in training duration, which may follow naturally if the foundational
knowledge in various fields expands as science advances.6 This idea can be examined in several
ways. First, Table 2 shows that training duration for Nobel Prize winners, measured as mean age
at Ph.D., increased by over 4 years over the 20th Century.7 The role of training duration can be
established more causatively by considering exogenous interruptions to young careers. Jones
(2010) employs World Wars I and II as such career interruptions and shows that these
interruptions must be “made up” after the war, producing both (a) unusual delays in the
completion of formal education and (b) unusual delays in the age of great achievement.
Furthermore, Jones and Weinberg (2010) show that the age dynamics in great
achievement within Nobel fields closely mirror field-specific dynamics in Ph.D. age. Generally,
for Nobel Prize winning work performed prior to 1900, 3 of 4 prize winners had received their
Ph.D. by age 25. For Nobel Prize winning work performed since 1980, only 1 of 5 prize winners
had a Ph.D. by age 25. Jones and Weinberg (2010) further analyze the effect of an exogenous
shock to the foundational knowledge in a field, studying the age and training patterns around the
quantum mechanics revolution in physics. The quantum mechanics revolution is typically
charted between 1900 and 1927 (e.g. Jammer 1966). Remarkably, we find that (a) age at great
achievement and (b) age at Ph.D. actually declined in physics during this period, reaching a
minimum just as quantum mechanics becomes a rigorously established theory in the late 1920s
and then rising thereafter. Moreover, these patterns are unique to physics; the age of great
achievements and Ph.D. age in other fields continued to rise during this period. Viewed as a
6 By contrast, a Kuhnian revolution in science may be associated with a contraction in the knowledge space,
temporarily reducing training requirements. See the discussion of the quantum mechanics revolution below.
7 Age at Ph.D. is a noisy delimiter of the boundary between a focus on training and a focus on active innovation.
That the Ph.D. age trend is somewhat smaller than the trend in age at first patent (an output-oriented delimiter) or
age at great achievement suggests that other intermediate institutions, such as the rise of post-doctorates, as well as
leaning-by-doing in the innovative process or other features, may involve further delays.