From Measuring Desire to Quantifying Expectations: A Late Nineteenth-Century Effort to Marry Economic Theory and Data - "Raw Data" Is an Oxymoron - page 57

collected and presented in the first place, even as his use of them effaced their historical
and conventional nature.
Conclusions
Now we can see why Fisher abandoned the mechanical-analogical method featured in
his earlier work and why this change in representational mode has proved to be so
consequential. The util, which is the unit of measurement he created to facilitate com-
mensuration in his analog-machine method, could only represent economic processes
analogically—only, that is, by producing a simulation that reproduced the theoretical
assumptions formulated as equilibrium theory. This method could not quantify desire,
in other words. What the analog actually depicted was simply the theoretical assumption
with which Fisher began: in an idealized economy, the equilibrium between desire/
demand and supply establishes prices. The mode of representation Fisher adopted in
Appreciation and Interest , by contrast, could refer to real-world transactions because the
prices he quoted were those actually paid by market participants and because the algo-
rithm he used to express his theory was the same algorithm used in those transactions.
Indeed, this mode of representation was simultaneously mathematical, empirical, con-
ventional, and theoretical: In marrying theory to empirical data in a quantitative form,
Fisher created a framework that could fit the theory to the data—but only if various
kinds of data were made commensurate with each other and only if the data embedded
the assumptions that made them useable in the first place.
The labor involved in rendering the available data commensurate, which Fisher
describes in the footnotes to Appreciation and Interest , might seem to undermine even
the tentative conclusions Fisher presents in this work. But it is important to acknowl-
edge how novel Fisher's ambition was in the late nineteenth century. Almost all of his
peers in the profession, even those who celebrated empiricism and fact-based work,
tended simply to gesture toward what data might prove, if it were to ever to be avail-
able. 25 AsWilliam J. Baumol has tartly remarked, the method of almost all late nineteenth-
century American economists relied primarily on “categorical pronouncements based
on no more than personal conviction.” 26 Fisher's data might seem to us to be both
uneven and unduly laundered, but his determination to use what data existed and his
painstaking descriptions of how he made it useable foretold the directions economics
would eventually take.