been consulted only by a few other scholars for
previous research (e.g., Crabbé 1986; Darnell
1988, 1990; Hands and Mirowski 1998), but
none of these works offer a comprehensive view
on Hotelling’s project regarding the economics
of non-renewable resources. The documents
stored at Columbia University consist of several
thousand sheets, filed in 58 boxes. Items related
to exhaustible resources are scattered in different
boxes. Thanks to a systematic review, we found
about 20 to 30 significant unpublished drafts,
preparatory documents and letters from 1924 to
1925 onwards worthy of interest for our inquiry.
4
Our paper is organized as follows. In
section 2, we investigate what sort of research
Hotelling was conducting in the 1920s, showing
his hesitation between a theoretical abstract
ambition and empirical concerns for what he
named at the time exhaustible resources.
5
This
section allows a clearer picture on the area of
validity, from the outset, for his basic result—
which is different than what has been retained
from his article. In section 3, we closely review
Hotelling’s work on geological constraints for
his 1931 article. We reveal a neglected
fundamental role for them in his analysis.
Hotelling considered cumulative production x(t)
and the change in production q’(t) as
fundamental variables, and not additional
factors, constituting a potential alternative basic
model capable of attaining U-shaped trajectory
for prices—the now empirically accepted
stylized fact for non-renewable resources. In
section 4 we provide our concluding remarks,
reassessing the Hotelling rule as the starting
point of mineral resource analysis.
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
4
This paper is part of the research project “Bifurcations in
Natural Resource Economics” (BNRE), sponsored by the
European Society for the History of Economic Thought
(ESHET). The objective has been to explore the genesis of
modern natural resource economics, notably through the
creation of a digital database of Hotelling’s economic
archives (7000+ images taken of 3000+ documents). The
project team mobilized researchers in natural resource
economics and history of economic thought over 2.5 years,
producing to date several coordinated articles and working
papers on complementary aspects related to Hotelling’s
research (e.g., Franco et al. 2019, Gaspard and Missemer
2019, Missemer et al. 2020, Missemer and Nadaud 2020).
5
What Hotelling reffered to as “exhaustible resources” are
now called “non-renewable resources.” In the following, we
use the modern term, except when it comes to evoking and
analyzing Hotelling’s project in its historical context.
2. Hotelling’s research on exhaustible
resources in the 1920s
2.1. From (natural) resources to assets, and
vice versa
Until the submission of his paper to the
Journal of Political Economy, most likely in the
fall of 1930,
6
Hotelling’s archives reveal,
regarding exhaustible resources, that he actually
worked along two approximately parallel lines of
research: a generic project on the valuation of
exhaustible assets, in the line of his theory of
depreciation, and a specific project on
exhaustible natural resources (minerals, fossil
fuels), more empirically grounded. In his final
article, Hotelling uses several terms to describe
his research: “exhaustible (natural) resources,”
“exhaustible assets,” “irreplaceable (natural)
resources,” “irreplaceable assets,” “mineral
content,” “resources of the earth,” etc. By
definition, the set of “natural resources” here
contains both fossil energy sources (coal, oil,
gas) and minerals. The set of “assets” is broader,
containing “natural resources” as a subset. A
mineral stock is an asset. A financial bond is also
an asset, but obviously not a natural resource.
Focusing on “natural resources” suggests that the
precise subject of research is related to concrete
forms of natural items. On the opposite side,
focusing on “assets” opens the reflection to other
case studies, beyond natural items.
7
Hotelling’s drafts, correspondence and notes
throughout the 1920s confirm the co-existence of
two levels of analysis: a generic level, illustrated
by the term “exhaustible assets,” and a more
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
6
Two letters from Hotelling to R. A. Fisher and
Chamberlin (dated July 23, 1930, and Aug. 8, 1930,
respectively) indicate that Hotelling was working on the
final version of the paper in the summer of 1930 (RAFP,
digitized version online; Guicherd 2017, p. 179). A letter
from Mills (dated Jan. 17, 1931) shows that the final
version of the paper was already circulating at the very end
of 1930 (HHP, Box 1, Mills, Frederick C.). Arrow (1974,
p. 1103; 1980; 1987, p. 670) reports that Hotelling (who
was his PhD supervisor) submitted his paper to the
Economic Journal before submission to the Journal of
Political Economy. As shown in Gaspard and Missemer
(2019), there is, however, no clear evidence that he did so.
7
This statement is not retrospective. In the early 20th
century, Irving Fisher (1906) had already discussed the
generic nature of the words “assets” and (not necessarily
natural) “resources.”