grants. To maintain neutrality with other careers, a simple rule of thumb is as follows. If
training duration rises by Y years per decade, then with an R% discount rate, V would need to
increase RY% per decade relative to the value of alternative careers that do not feature extending
training phases. For example, the evidence surveyed in Section II suggests that Y averages about
0.8 years per decade. With a 10% discount rate, V would need to increase by 8% per decade
beyond the value increases of other careers. A simple way to achieve this would be to increase
salary support by 8% per decade or 0.8% annually above real wage growth in those careers with
static training. A closely related alternative would be to increase wage support during the
training phase, through graduate student and post-doctoral stipends.13
One might also increase V through other dimensions. For example, longer, larger, and/or
less restrictive research grants at the height of the scientist’s career may be attractive in
expectation and help offset the automatic disincentives that emerge as training duration
increases. At the same time, forcing grant dollars (not wage support) earlier in the life-cycle
looks sub-optimal, in the sense that early-life cycle researchers are less likely to produce
important ideas, as shown in Section II.
An additional alternative is to accelerate training. This approach may be especially
attractive and of increasing importance if an individual’s raw innovation potential is greatest
when young. Historically, Figure 2A suggests high innate innovation potential among young
scholars (i.e. were training not occupying the individual’s time), which is consistent with the
broader literature on life-cycle creativity.14 This finding further amplifies the opportunity costs in
the early life-cycle and especially the costs of “busy work” professional apprenticeships, where
future innovators are saddled with rote, relatively low skill tasks that have little training value.
As one response, science policy might increasingly emphasize a separate track of professionals
who focus on rote analytical tasks, requiring less training and without anticipation of being
13 This discussion emphasizes keeping entry incentives “neutral” with respect to other careers. Of course, one may
imagine that research support levels are too low or high in general, and neutrality is meant only as one benchmark.
14 The capacity for great ideas from young scholars is shown historically in Figure 2A when considering the 1900
estimate. See also Stephan and Levin (1993), Simonton (1998), Weinberg and Galenson (2008), Jones (2010) and
Jones and Weinberg (2010) for further discussion.
21