is owned and operated by the Economic History Association
with the support of other sponsoring organizations.

Marshall’s Tendencies: What Can Economists Know?

Author(s):Sutton, John
Reviewer(s):Larson, Bruce

Published by EH.NET (March 2002)

John Sutton, Marshall’s Tendencies: What Can Economists Know? Cambridge,

MA: MIT Press, 2000. xvi + 122 pp. $22.95 (cloth), ISBN: 0-262-19442-2; $14.95

(paperback), ISBN: 0-262-69279-1.

Reviewed for EH.NET by Bruce Larson, Department of Economics, University of

North Carolina at Asheville.

John Sutton teaches microeconomics and industrial organization at the London

School of Economics and Political Science. In his previous books, Sunk Costs

and Market Structure and Technology and Market Structure, he used

game theory and history to explore a range of industries from salt to sugar and

from coffee to beer, as well as to develop the “bounds approach” to market

structure. In Marshall’s Tendencies , which is based upon the Gaston

Eyskens Lectures that Sutton gave at the University of Leuven in 1996, he

builds upon his previous work, while also providing a valuable perspective on

the emergence of the standard paradigm in economics and what other economists

might do. In relation to this, Sutton asks: “Is it possible to find economic

models that work?”

The epigraph of Sunk Costs and Market Structure, taken from Joseph

Schumpeter’s History of Economic Analysis, provides a useful framework

for Sutton’s work. It states:

There are such things as historical and theoretical temperaments. That is to

say, there are types of mind that take delight in all the colors of historical

processes. . . . There are other types that prefer a neat theorem to everything

else. We have use for both. But they are not made to appreciate one another.

The tension between these temperaments is clear. It is recognized, however,

that an appropriate division of scholarly labor might necessitate that those

with exclusively historical temperaments focus on history and those with

exclusively theoretical temperaments focus on theory. Still, there are those

whose temperaments combine the historical and the theoretical, and there is a

use for them, too. Such a person is John Sutton and such a person was Alfred

Marshall. Such people will presumably seek to find an appropriate blend of

theory and history in their work.

Marshall’s Tendencies contains four chapters. In chapter 1 (“The

Standard Paradigm”) Sutton discusses Marshall’s analogy of the tides, first

introduced in the third edition of Principles of Economics. In contrast

to planetary movements, which can be predicted very accurately, tidal movements

are hard to predict, for although both phenomena are primarily affected by

gravity, the tides are secondarily affected by meteorological factors. It is

due to these seemingly random secondary factors that tides are so hard to

predict and it is because economic life is subject to so many secondary factors

that economic laws must be expressed as tendencies.

Sutton notes that Marshall’s distinction between primary factors and secondary

factors was given its first systematic use in the work of Trygve Haavelmo, in

which the primary factors comprise the “systematic part” of the econometric

model and the secondary factors are reflected in the “disturbance term.” This,

combined with the comparative statics of Paul Samuelson, effectively comprises

the standard paradigm. The thrust of chapter 2 (“Some Models That Work”) is to

show that the standard paradigm works well when there is no problem of model

selection, that is, when the true model is known, as in option pricing and some


But what if the true model isn’t known? What if the problem of model selection

is intractable? That is the subject of chapter 3 (“Relaxing the Paradigm”). Its

argument unfolds in relation to the nineteenth century thermodynamics of Sadi

Carnot, whose work progressed when he greatly increased the level of its

abstraction. In relation to this Sutton indicates that even if one cannot write

down a complete model of some phenomenon, one might still be able to develop a

theory that generates testable hypotheses. The “bounds approach” exemplifies

this. As he observes with respect to the usual cloud-like relationship between

concentration and R&D intensity, “Using a bounds approach . . . we arrive at a

clear and sharp characterization” (p. 84). Sutton is careful to note, however,

that the bounds approach should be seen as complementary to the standard

paradigm and not as a substitute for it — the bounds approach bears fruit in

cross-section studies because all of the complicating factors that relate to

them operate in the same direction (p. 85).

The bounds approach, then, provides a means for developing testable hypotheses

beyond those provided by the comparative static methods of the standard

paradigm (please see Hale, et al., for an overview of these methods). Sutton

draws upon this in chapter 4 (“Testing”). While recognizing the late-twentieth

century worries about theory testing and the pessimism which resulted from

them, he indicates that there are economic models that work and with

imagination more can be developed. In relation to this Sutton simply notes: “If

we take such worries seriously, we will adopt a relaxed and eclectic view as to

research methods, and we will be the better for doing so.” This is a view with

which Marshall would have surely agreed.

Marshall’s Tendencies builds upon John Sutton’s previous work, while

also providing a valuable perspective on the emergence of the standard paradigm

in economics and what other economists might do. It is an extremely stimulating

book worthy of the consideration of anyone who is interested in the potential

(and limits) of economic knowledge. His characterization of the emergence of

the standard paradigm is broadly correct and his discussion of theory

performance in relation to model selection rings true. Furthermore, his bounds

approach indicates one direction in which future work might be fruitfully done.

Finally, his example shows that, as paradoxical as it might seem, empirical

advances can sometimes be made through greater abstraction. Sutton’s work

clearly shows that an artful blending of theory and history can take us places

inaccessible solely by theory or history, but it also indicates that further

advance is unlikely to be easy. Few things worth doing ever are. References:

Douglas Hale, George Lady, John Maybee, and John Quirk, Nonparametric

Comparative Statics and Stability. Princeton, NJ: Princeton University

Press, 1999.

John Sutton, Sunk Costs and Market Structure: Price Competition,

Advertising, and the Evolution of Concentration. Cambridge, MA: MIT Press,


John Sutton, Technology and Market Structure: Theory and History.

Cambridge, MA: MIT Press, 1998.

Bruce Larson studies the history of mathematical economics, his most recent

article being “Canard on Direct Exchange and Taxation: A Perspective on

Cournot,” History of Political Economy 31, no. 1 (Spring 1999): 109-31.

Subject(s):History of Economic Thought; Methodology
Geographic Area(s):General, International, or Comparative
Time Period(s):General or Comparative