Economic Interests and Threat Assessment in the U.S. Congress, 1890-1914
Abstract
Why do some domestic actors see the international environment as a threatening place populated
by untrustworthy powers, when others view the same conditions in a more benign light, finding
opportunities for peaceful cooperation? Because international conditions do not vary across
domestic political actors, the answer to this question must lie in the traits of those evaluating
them. In this paper, we consider the impact of economic interests that might link threat
assessment to broader societal cleavages. We examine evaluations of the international
environment in speeches given in the U.S. Congress during naval appropriations debates between
1890 and 1914. In addition to its historical importance, this period provides an opportunity to
observe threat assessments in an emerging power when the weight of past policies and
commitments is relatively light. The manufacturing sector's interest in trade protection led
political leaders who represented manufacturing regions to offer more negative assessments of
the international environment, while those representing export-oriented agricultural areas of the
country gave more positive evaluations. These effects were roughly comparable to those
associated with individual-level characteristics, such as having served as a military officer.
Michael E. Flynn
Benjamin O. Fordham
Department of Political Science
Binghamton University (SUNY)
Research for this paper was supported by the National Science Foundation Political Science
Program through grant SES-1022546. The authors would like to thank participants in the World
Politics Workshop at Binghamton University, the Buffett Working Group on Security Studies at
Northwestern University, and the Departments of Political Science at the University of Alabama
and Michigan State University for their comments and suggestions. All errors and omissions are
our responsibility.
To decline to fortify [the Panama Canal], in my opinion, opens broad the way of
peace and preserves friendly relations with all other powers. To fortify means we
are looking for war; and, looking for it, we will sooner or later find it.
--Representative William Richardson (D-AL)
As much with the nation as with a man, the first law in nature, and therefore its
first duty, is self-preservation. The law of self-preservation dictates that this
Nation shall fortify the canal and control it when questions of national defense
arise.
--Representative Richmond Pearson Hobson (D-AL)
On February 25, 1911, the House of Representatives spent much of its floor session debating a
measure to construct seacoast batteries at the entrances to the Panama Canal, which was then
nearing completion. 1 Some members, such as Representative Richardson, quoted above, argued
that fortifying the Canal would be unnecessary and provocative. They preferred an international
agreement to respect the neutrality of the Canal instead. Moreover, they pointed out that the 1901
Hay-Pauncefote Treaty, in which the British government conceded the American right to control
an isthmian canal, provided that the canal remain neutral and open to the ships of all powers in
both peace and war. Other members, such as Representative Hobson, scoffed at the idea of a
neutralization agreement. As Hobson put it, "[m]ust we depend for the security of our vital
interests upon the fickle pledges of other nations?" They showed little concern about treaty
obligations, arguing that the United States should seize whatever military advantages the Canal
might provide in time of war. Though the House ultimately accepted the fortification provision
by a vote of 123-81, the debate revealed sharp differences over the nature of the international
environment, the value of international agreements, and the need to forcefully assert American
interests over those of other states.
1
Both quotes are from page 3449 of the Congressional Record for February 25, 1911.
1
Debates like this one have taken place in many nations and in many historical periods.
The politics of foreign policy frequently involve conflicts among domestic factions with very
different views of the international system. Some hold that it is a threatening place, populated by
predatory powers that cannot be trusted to keep their commitments. Others view the same
international conditions in a more optimistic light, seeing opportunities for mutually beneficial
cooperation. What accounts for these different assessments? Because all observers confront the
same international environment, the conditions themselves cannot provide a complete answer.
When it comes to explaining differences among policymakers, individual psychological or
ideational processes also tell only part of the story. A range of other considerations also influence
their threat assessments, including prior policy commitments that have to be justified in terms of
international threats, their cooperative arrangements with other policymakers, and the interests of
their constituents.
In this paper, we will focus primarily on the last of these considerations—constituent
economic interests. We argue that these interests lead political leaders to take different positions
about whether international conditions contain mainly threats or opportunities for cooperative
engagement. We test this argument by examining the politics of foreign policy in the United
States between 1890 and 1914, when the nation first began to assert itself in great power politics.
During this period, the manufacturing sector's interest in trade protection led those representing
industrialized parts of the country to see the world in a more threatening light than did those
representing export-oriented agricultural regions. This period provides a good opportunity to
observe the processes that shape threat assessment in an emerging power under conditions where
the weight of the past is relatively light. American leaders were just beginning to define the
nation's overseas interests. Compared to later periods, prior policies and commitments did not
2
provide strong guidance about the implications of the nation's growing significance in world
affairs or its relations with other major powers. The United States also lacked a well-developed
foreign policy bureaucracy that might have transmitted a conventional wisdom about
international conditions to elected leaders. For example, the State Department had only 76 U.S.based employees in 1890, and only 234 in 1910 (Trask 1981, 42).
The evidence we consider here comes from a historical period that ended roughly a
century ago. Some scholars will find this period interesting for its own sake, as a source of data,
or as a spur to theoretical development. Still, those who are more concerned about explaining
present-day political phenomena or predicting future outcomes might question its relevance.
Why not instead develop arguments tailored to explaining contemporary outcomes and test them
using more recent data? These questions are important. Some of the broader patterns and
theoretical processes we will consider here are immediately relevant to these scholars' concerns,
but the different historical context does imply that not all of the specific relationships we discuss
map neatly onto present-day politics. Nevertheless, the past still deserves our attention. Social
science theories seek to explain important phenomena in the most general way possible, yet most
empirical tests use data drawn exclusively from the post-World War II period, reflecting an
international landscape that is unusual in many ways. Though scholars often assume their
theoretical arguments apply to earlier eras, their relevance might actually hinge on conditions
that are part of the common sense today but that did not hold in earlier periods. By limiting our
analysis to a time period that is often neglected, or simply pooled with other time periods, we can
better evaluate how generalizable our theories are. Inconsistencies, should any be uncovered, can
pave the way for further theoretical development and can shed light new questions for scholars to
explore.
3
The remainder of this paper proceeds in four parts. The first will explain what we mean
by "threat assessment" and outline our argument about the role of economic interests in shaping
it. The second explains the application of our argument to the United States between 1890 and
1914. Third, we will test hypotheses drawn from our argument using evidence about evaluations
of the international environment in Congress during the 25 years before World War I. The final
section summarizes our results and their implications for similar processes in other historical
periods.
Trade, Protectionism, and Threat Assessment
Our principal interest in this paper is in policymakers' assessments of the international
environment as either threatening or benign. As we will explain in more detail below, we will
use statements members of Congress made during naval appropriations debates as a source of
data on these evaluations. We will refer to these types of evaluations as "threat assessment" to
distinguish them from the more commonly discussed phenomenon of "threat perception." Most
previous work exploring threat perception at the sub-national level has emphasized individual
interests, psychological processes, and ideological predispositions. 2 Empirical analyses have
typically employed surveys of the general population (e.g., Rousseau 2002; Rousseau and
Garcia-Retamero 2007; Fordham and Kleinberg 2010; Kleinberg and Fordham 2011) or groups
of college students in a laboratory setting (e.g., Kemmelmeier and Winter 2000). Although they
offer several different explanations, these studies are all concerned with immediate individual
reactions to the international environment.
2
Of course, research treating states as unitary actors also frequently employs the concept of international threat (e.g.,
Walt 1987) but this work abstracts away from the sub-national differences in evaluations of international conditions
that concern us here.
4
At first glance, policymakers' statements about the international environment might seem
to be another instance of threat perception, similar to responses to survey questions about
international conditions. In fact, threat perception, as most research on the topic understands it, is
only one component of the elite threat assessment we will consider here. A range of concerns
apart from their own individual perceptions of the international environment influence
policymakers and might lead them to take positions that differ from those they would have
expressed if they were instead responding anonymously to a survey question. Most obviously,
policymakers' positions on international conditions have consequences that those of survey
respondents do not—their opinions not only directly affect policy outcomes, but can also shape
public opinion on vitally important matters such as war and peace (e.g., Berinsky 2009; Howell
and Pevehouse 2007, 155-221). Policymakers thus have reason to be cautious in their public
evaluations of international conditions, even if doing so leads them away from their own
individual perceptions and beliefs.
Another important difference between threat assessment by elites and threat perception in
surveys of the general public is that policymakers must consider the interests and opinions of
their constituents when evaluating threats and opportunities. Indeed, their constituents'
immediate economic interests may be more important than their own. Some policymakers might
share their constituents' interests through investments or property ownership in their home area.
However, most draw their income from the state and are thus are more insulated from the
immediate effects of international trade than their constituents are. To the extent that politicians
internalize their constituents' interests, this distinction may make little behavioral difference.
However, the fact that they could potentially build a reelection coalition around other issues or
around constituents who have different economic interests places them in a somewhat different
5
position than individuals contemplating their own interests. This process of representation is our
primary concern in this paper.
How could societal economic interests affect threat assessment? Our answer builds on
previous research linking these interests to foreign policy preferences on many issues. Trade is
one of the most important forms of contact between states, so it is not surprising that its effect on
their constituents influences policymakers. A wide range of international conditions and policy
choices could disrupt access to international markets, so many foreign policies that are not
immediately concerned with trade nevertheless affect it. For example, Milner and Tingley
(2011), Cutrone and Fordham (2010), and Fordham (2008) have connected trade policy interests
to congressional votes and speeches on foreign aid, human rights, and national security issues,
respectively. In each case, the relationship arises from the fact that policies on these issues can
impinge on trading relationships that ultimately affect constituents' income.
We wish to test the linkage between preferences for trade protection and assessments of
the international environment. Briefly stated, we expect policymakers whose constituents stood
to benefit from trade protection to identify greater threats from other states and a more hostile
international environment in general. Alternatively, those whose constituents preferred open
international markets should be less apt to assess the international environment in such a
negative way. As other scholars have pointed out, trade protection often implies a hostile foreign
policy. For example, Patrick McDonald (2009) presents evidence that states dominated by
protectionists are generally more conflict prone. While those who want to uphold or expand
commercial relationships with other states have a stake in avoiding political and military conflict
with their trading partners, protectionists have no such incentive. Indeed, they might actually
benefit from hostile relations. Even if protectionists would prefer to avoid political hostility with
6
trading partners, the fact that their commercial policy inflicts material harm on these states may
make that difficult.
Both liberal and protectionist commercial policies have broader foreign policy
implications because the basic conditions for exchange across international borders are not
always secure. However, the foreign policies that complement protectionism are especially likely
to create conflict with other states. For example, Eckart Kehr (1977 [1927], 22-49) argued that
German hostility to both Russia and Britain during the years before World War I stemmed from
protectionist policies that different elements of the ruling coalition in Germany demanded.
Grain-growers in the East and industrialists in the West saw Russia and Britain, respectively, as
key competitors. In Kehr's account, high tariffs and complementary policies, such as interference
with Russian access to capital, the construction of a battleship fleet, and the pursuit of exclusive
access to overseas markets in less developed areas, made both of these states into enemies.
Sophisticated protectionists—knowing that their opportunistic commercial policy make political
hostility more likely—should expect other states to behave in this way and should advocate an
aggressive foreign policy posture from the beginning. Less sophisticated protectionists might
simply lack any motive to cooperate with other states and view the negative responses to
protectionism as evidence of their hostile intentions.
For their part, export-oriented interests that stand to benefit from a more liberal trade
policy have no reason to support either trade protection or the aggressive foreign policies that
often accompany it. Because of the size of their economies, the developed states that
protectionists view as their primary competitors are often also the most promising markets for
these exporters. Accordingly, they might come to view these developed states as potentially
valuable friends with whom cooperation in the pursuit of joint gains is a real possibility.
7
Conversely, hostile political relations with these states, as well as their likely retaliation against
the high tariffs protectionists seek, threaten the interests of many exporters. Even setting aside
the direct harm it causes to their interests, political-military competition with other states is
expensive and dangerous. Policies protectionists see as necessary preparations for inevitable
political conflict might strike free traders as expensive, provocative, and unnecessary.
Constituent economic interests of this sort could be transmitted to policymakers in many
different ways. Much of the work on this topic suggests that constituents can offer votes,
campaign contributions, and other political resources in order to influence their political
representatives. This exchange of resources for policy positions is a useful theoretical
simplification, but probably does not accurately describe the way the process works in many
cases. In fact, members are likely to hold policy views that match the interests of their
constituents because they formed these views through interactions with those constituents rather
than as a result of a straightforward exchange. Political leaders might simply care about the
economic welfare of their home area and seek out policies that will promote it. They might have
formed their views alongside their constituents in a political and social milieu strongly colored
by the economic interests prevailing in their home area. Constituent interests could influence
policymakers through many other subtle pathways that are difficult to disentangle analytically.
Constituents are not the only group that might influence policymakers' threat
assessments. Legislators and other policymakers must also interact extensively with one another
in order to achieve their goals. Partisan affiliations and common ideological orientations reflect,
among other things, formal and informal cooperative arrangements for this purpose. These
considerations might have independent effects on policymakers, possibly pulling them toward
the positions of other policymakers and away from what they or their constituents might prefer.
8
In this sense, party and ideology compete with constituent interest to influence policymakers. At
the same time, though, their constituents will use these affiliations to select political
representatives whose broad ideological positions are consonant with their interests. Such
representatives are sure to be more sympathetic to their interests when constituents bring relevant
policy issues to their attention. To the extent that constituent interests affect the party or ideology
of those chosen to represent them, the effect of party and ideology on the positions these
representatives later adopt is actually due to those underlying constituent interests.
Constituents' use of party or ideology to select representatives sympathetic to their
interests is an important issue in assessing the influence of these interests on policymakers.
Fordham and McKeown (2003, 523) distinguish between a "selection" and an "influence" effect
of constituent interests. Selection refers to influences on candidate recruitment, nomination, and
election. Influence refers to efforts to affect the votes of sitting legislators. If their impact on the
selection process is sufficiently large, constituents might have no need to lobby sitting
legislators. Making a parallel case using the language of principal-agent theory, Bearce (2003,
382) argues that societal interests, acting as principals, select representatives to serve as their
agents based on the ideological commitments that their party affiliation implies. The societal
constituents need only lobby if their representative shirks by taking positions that run against
these ideological commitments. However one frames the argument, it suggests that at least some
of the effect of constituent interests on legislators' views of the international environment should
work through the election of ideologically sympathetic candidates rather than changes in the
positions of those already holding office. While it is not practical to disentangle the many ways
that constituents might directly influence policymakers' positions, the selection process must be
9
considered in order to accurately assess the effect of constituent economic interests on threat
assessment. We will return to this issue in discussing our research design.
To be clear, our claim is not that domestic economic interests are more important than
international conditions. Rather, it is that the economic interests of domestic actors shape
whether international conditions pose threats or present opportunities to them. As we will explain
the case of the United States before World War I, domestic and international considerations
interact to shape the politics of foreign policy. This line of argument does not imply that there
will always be political conflict over threat perception or national security policy. Some events
and conditions pose a threat to a wide range of economic interests, pushing most or even all
domestic factions toward a shared evaluation of the international environment, and perhaps a
common set of policies to confront it. Moreover, economic interests are clearly not the only
considerations that shape the politics of foreign policy. The psychological and ideological factors
discussed in the literature on threat perception noted earlier might make a difference. These and
perhaps other considerations might overcome the policy differences suggested by divergent
economic interests. The actual influence of economic interests on threat assessment is an
empirical question.
In order to apply this rather abstract discussion of economic interests and threat
assessment to the United States in the 1890–1914 period, some knowledge of the historical
context is necessary. As we will explain in the next section, American manufacturers were
strongly committed to trade protection during this time period, while agriculturalists stood to
benefit from more liberal trade policies. The foreign policy debates of the time also reflected the
fact that the United States was only one power among many and lacked both the prestige and the
leverage over other developed states that it would have after 1945. It is also significant that the
10
world was not yet divided by the overarching ideological conflicts that would emerge later in the
20th Century. These and other considerations were critical to the politics of foreign policy during
the period we will examine but prevent the simple extension of the patterns evident before World
War I to later historical periods. We will return to the matter of historical context in our
conclusion.
The Economic Stakes in the American Rise to World Power
The United States emerged as a state potentially powerful enough to influence the shape of the
international system in the late 19th Century, but there was no broad agreement about how the
nation should use its newfound power. As this section will explain, two factions with very
different views of the international system contended for control of American foreign policy.
One, associated with manufacturing interests and the Republican Party, sought access to markets
in less developed areas of the world while protecting the American domestic market against
manufactured imports. This policy brought the United States into conflict with other major
powers who were seeking to do much the same thing. The other faction, associated with
agricultural interests and the Democratic Party, preferred to maintain cooperative relations with
the other major powers, who were the primary markets for American agricultural exports.
The political economy of American foreign policy between 1890 and 1914 begins with
the different overseas interests of manufacturing and agriculture. American manufacturing
expanded rapidly during the 19th century. Maddison (1992) estimates that the United States
surpassed Britain to become the largest economy in the world by 1872. As the map in Figure 1
suggests, manufacturing was largely confined to the northeastern quadrant of the United States
during this period. The 15 states between New England and the Great Lakes accounted for 77
11
percent of manufacturing employment and output in 1900 (United States Census Bureau 1901).
The rest of the country remained largely agricultural. Thus the interests of manufacturing were
the interests of only one region of the country, albeit a very populous and important one.
*** Figure 1 about here ***
The divergent interests of manufacturing and agriculture were nowhere clearer than on
the tariff, one of the most important political issues of the day. Demands for trade protection
came mainly from manufacturers, though producers of wool and sugar also played a role. From
the Civil War onward, they obtained relatively high tariffs on manufactured imports. By contrast,
American agricultural commodities were highly competitive and constituted the bulk of
American exports before World War I. Figure 2 illustrates this pattern. Though commodities like
wheat and tobacco were also important, cotton was by far the largest American export,
comprising roughly one quarter of all exports from the end of Reconstruction through 1914.
Manufactured exports grew during this time period, but remained far less important than
agriculture. Combined exports of iron and steel, machinery, and automobiles never constituted
more than 10 percent of American exports until 1910, and were well below this number for most
of the period. Under these circumstances, the tariff transferred wealth from farmers to
manufacturers. As one scholar explained in the North American Review, "[o]ur farmer sells on
the basis of Liverpool prices in the market of the world. He buys home productions at a protected
price, and thus he is hit both ways" (Selbie 1893, 254).
*** Figure 2 about here ***
12
The tariff controversy lay at the root of partisan conflict over foreign policy. The
Republican Party was strongly committed to trade protection, especially after 1887 (Goldstein
1993, 19-20; Epstein and O'Halloran 1996, 301-24; Taussig 1910, 216-9). This commitment was
as clear to political actors at the time as it has been to subsequent scholars. Between the elections
of 1892 and 1912 the Republican Party Platform consistently endorsed the tariff as a core
component of the party's agenda (Republican National Committee 1892; 1896; 1900; 1904;
1908; 1912). Republican presidents could not easily depart from this position. As Bensel (2000,
457-509) explains, tariffs provided a vital incentive for key constituencies to support other
aspects of the Republican Party's program. More than 99 percent of Republicans in Congress
voted for trade protection between 1888 and 1897. Less than 4 percent of Democrats took this
position. Predictably, Democrats from agricultural states were most strongly opposed to the
tariff.
Although tariff reformers had a few successes along the way, advocates of trade
protection had the upper hand politically. The nation's population, like its manufacturing
industries, was concentrated in the Northeast. The 15 manufacturing states in the Northeast and
Great Lakes region held 214 of the 224 electoral votes needed to win the presidency in the 1900
election and were similarly critical in all the other elections during the 1890-1914 period (U.S.
National Archives and Records Administration 2012). Democratic candidates, who typically
endorsed "tariff reform," needed to win at least a few of these manufacturing states in order to
gain the White House. They accomplished this difficult task only twice in the seven elections
between 1888 and 1912. By contrast, advocates of trade protection could assemble a successful
electoral coalition by gaining only a few agricultural states, something they sought to do by
13
protecting domestic producers of wool and sugar in addition to manufactures. Republicans
controlled the White House continuously from 1896 through 1912 and were thus able to
decisively shape American foreign policy during the period considered here. Their agenda was
the principal one debated in Congress.
The predominance of protectionists had broader consequences for American foreign
policy that are not widely appreciated. Historians of the Wisconsin School have extensively
documented elite demand for overseas markets in the late 19th Century (e.g., LaFeber 1963, 15096; McCormick 1967, 21-52; Williams 2009 [1959]). Most of this agitation centered on markets
in less developed areas of the world, particularly Latin America and East Asia. This regional
focus should appear more puzzling than it does in most historical accounts. Europe provided by
far the largest market for American exports during this period, constituting roughly 75 percent in
1900, compared to 16 percent for the Americas and 5 percent for Asia (Historical Statistics of the
United States 2006). Even in Asia and the Americas, the relatively developed markets of Canada
and Japan were the main destinations for American products.
These regional priorities only make sense when one considers the implications of trade
protection in the United States. Maintaining high tariff barriers, especially on manufactured
goods, provoked retaliation from European trading partners and complicated efforts to promote
American exports to developed countries. Less-developed trading partners exported few
manufactures and thus had little cause to complain about high American tariffs on these goods.
As one Republican policymaker put it when advocating trade with Asia, "the best hope of
American commerce" was "trade with people who can send you things you want and cannot
produce, and take from you in return things they want and cannot produce" (Reid 1900, 41-2).
14
The preference for less-developed markets mattered because maintaining access to these
markets required greater political and military activism than did trade with developed countries.
Internal political conflicts like the Boxer Rebellion in China disrupted trade and sometimes
prompted American military intervention. Such interventions were far more frequent in the
Caribbean Basin where American policymakers acted not only to protect American economic
interests but also to preclude similar actions by the European powers. The empire-building
efforts of these powers threatened American access to many less-developed markets. The
division of nearly the entire African continent into European colonies was an ominous precedent.
Preventing a similar colonization process in Latin America was a longstanding priority enshrined
in the Monroe Doctrine and enforced more aggressively after the 1895 Venezuelan Boundary
Crisis. After the acquisition of the Philippines, American policymakers also sought to prevent the
partition of China, most notably through the Open Door notes of 1899–1900. In short, the pursuit
of less developed markets brought the United States into political-military competition with other
major powers.
In view of other major powers' frank colonial ambitions during this period it is not
surprising that Americans interested in gaining access to less-developed markets commonly
sought this access on a preferential basis. Public discussions of American interests in Latin
American markets frequently turned to the need to "exclude European productions from our
Latin neighbors," as one manufacturer put it during a 1901 debate on reciprocal trade agreements
(National Association of Manufacturers 1901, 41). Important Republican policymakers such as
Henry Cabot Lodge (1895, 658) argued that the United States had a "rightful supremacy" in
Latin America and that its position in the hemisphere should be defended "peaceably if we can,
forcefully if we must." This was not mere rhetoric—Republican presidents sought to expand
15
access to Latin American markets through strictly bilateral reciprocity agreements. Adopting a
conditional interpretation of most-favored nation status, they refused to extend the tariff
concessions granted under these agreements to other states (Jones 1908; Viner 1924). Though
they were not always successful in getting Latin American trading partners to go along, they
tried to persuade them not to extend trade concessions granted to the United States to European
states (Taussig 1910, 240). Though the United States ultimately did not pursue further formal
colonial possessions after 1900, this course was attractive to some in part because the United
States could impose its preferred trade policy on colonies like the Philippines (Hoxie 1903).
Such an opportunistic commercial policy implied hostile relations and perhaps even
military conflict with other major powers. These states' foreign policies threatened American
access to less developed markets. Moreover, parallel American efforts to obtain exclusive
economic access, especially in Latin America, would inevitably inflict material harm on these
other powers. Among observers who favored trade protection or took its presence for granted,
commentary about the inevitability of political and military conflict over control of markets was
common. Brooks Adams (1900, 26-53) and Alfred Thayer Mahan (1987 [1890], 1) are perhaps
the best-remembered of those writing in this vein, but they were not alone. Proponents of
commercial expansion in less developed markets commonly discussed economic competition
using military metaphors such as "peaceful combat," "conquer[ing] by commerce," "peaceful
conquest" and the like (Eddy 1891, 421; Blaine 1892, 522; Carter 1894, 431). Much American
writing on international relations from the 1890-1914 period that assumed that all states would
protect their domestic market and seek privileged access to less developed markets overseas, a
situation many authors viewed as tragic but unavoidable. Not surprisingly, those who began with
16
this assumption tended to conclude that the international environment was necessarily
competitive and hostile. For example, Conant (1900, 608-9) noted that
[i]f commercial freedom were the rule among nations, so that there could be no
discrimination against the most efficient producer, the industries of the United
States would need no political support in the contest for commercial supremacy.
But, by reason of the condition which have prevailed in the world from the
beginning, under which diplomatic finesse and military force have been brought
to the support of national commerce, it is essential that those peoples who can
produce under the best conditions should not be deprived of the opportunity to
sell in the world's markets.
The same argument can be found in many textbooks from the time, such as Coolidge (1908, 17883), Latane (1907, 100), and Reinsch (1900, 38-9).
Then as now, advocates of free trade objected that trade was not a zero-sum endeavor.
Edward Atkinson (1896, 84) wrote during the dispute with Britain over the Venezuelan border
with British Guiana that "[t]he mediaeval error that in all commerce what one nation gains
another must lose, is one of the most potent causes of the animosity against England which now
depraves the minds of many people in this country." Protectionist Republican policymakers
preferred the militarized language, which sometimes threatened to become more than a
metaphor. As one harsh but not entirely inaccurate Democratic critic summarized their
perspective, "[i]ts spirit is that of conquest; its first reason, as well as its last, is force" (Gray
1895, 424).
Hostility was greater towards states that pursued a policy of protectionist imperialism
similar to that of the United States—a category that included essentially every major power but
Britain. Changing American relations with Japan illustrate the pattern. At the time of the Open
Door Notes, American observers identified Russia as the primary threat to American economic
access because of its newly acquired concessions in Manchuria (Hannigan 2002, 96-9). As
Brooks Adams (1898, 645) ominously put it "[t]he Russians hardly veil their purpose of
17
reversing, by means of railways, the current of the Chinese trade as it has flowed for ages, and of
using force to discriminate against maritime nations; but those who are excluded from the
Eastern trade have always lagged behind in the race for life." 3 Because Japan opposed Russian
ambitions, many Americans viewed it as a potential ally. Reviewing press commentary on the
Russo-Japanese War in 1904, the Literary Digest noted that "[n]ot a single newspaper in the
country, as far as we have seen, favors anything but absolute neutrality in the dispute between
Russia and Japan; but, at the same time, most of our press would like to see Japan win…With
Japan paramount, it is believed that American trade and influence in Asia would enjoy a rapid
growth" (American Sympathies in a Russo-Japanese War 1904, 101).
Once Japan defeated Russia and gained political control over much of Manchuria, it
quickly became a greater potential threat to American interests. American sympathies followed
the country's trade interests. The Literary Digest was soon summarizing articles citing Japanese
efforts to monopolize commerce in areas they had seized from Russia, calling for greater
American efforts to preserve Chinese independence against Japan, and pointing out Japan's quest
for commercial supremacy in Asia (e.g., Menace of Japanese Success 1905, 920; Our Duty to
Save China from Japan 1907, 288; Our Peaceful Contest with Japan 1907, 594). By 1908,
Forum's regular foreign policy commentator, Maurice Low (1908, 311) was writing that "[i]t
would be idle to attempt to disguise the feeling that so strongly exists in the United States.
Rightly or wrongly naval officers and public men believe that the United States has more to fear
from Japan than any other country, and Japan is equally suspicious."
3
Similar views from the same journal, Forum, include Low 1899; Cary 1898; Moffett 1899; and Reid 1901.
18
Testing the Influence of Economic Interests on the Politics of Foreign Policy
The qualitative evidence reviewed above suggests that demands from the manufacturing sector to
secure markets in less developed areas while maintaining high levels of trade protection against
manufactured imports from other developed states influenced elite discourse on foreign policy.
However, these examples alone do not demonstrate a systematic pattern in the politics of
American foreign policy. Even if they did, questions would remain about the importance of
economic interests compared to other influences. In this section, we will set out a quantitative
research design for addressing these questions using data on congressional deliberations over
foreign policy.
We use several types of data to test hypotheses about the influence of economic interests
on congressional assessments of the international threat environment. First, we measure each
state's economic structure using occupational data from the 1900 Census. For each state, we use
the proportion of workers in manufacturing occupations to indicate the size of the manufacturing
sector. These are the data reported in Figure 1. Though we have census data only on states rather
than congressional districts, this should not bias our results in favor of our claims about the role
of economic interests. If anything, the measurement error associated with state-level data should
bias against support for our hypotheses. In principle, it would make sense to include a separate
variable indicating the size of the agricultural sector. However, because manufacturing and
agriculture dominated the economy as a whole during this time period, these two variables are
essentially mirror images of each (r = -0.94). While we will use the size of the manufacturing
sector in the analysis that follows, the results should be understood to reflect the contrasting
interests of agriculture and manufacturing, rather than those of manufacturing alone.
19
Second, we use data on speeches given during naval appropriations debates to indicate
the foreign policy views of members of Congress. We chose these debates because they occurred
on a regular basis in every congress, not just when special domestic or international
circumstances required them. The Navy was the principal power projection tool available to the
United States during this period. On the other hand, the Army was small and had never been
deployed overseas in 1890. Consequently, the Army appropriations bill rarely engaged larger
questions about the international environment. Debates about the Navy regularly confronted
these issues throughout the 1890–1914 period. The size of the battleship fleet—the most
common bone of contention in these debates—depended critically on the intentions of other
states and the character of international politics. These debates are especially interesting because
members were essentially free to say whatever they liked. By contrast, their roll-call voting
record is more likely to reflect the efforts of the congressional leadership to control the agenda
and get members to follow the party's preferred position.
We coded both specific evaluations of other states and broader positions concerning the
international threat environment in general. We examined all the statements members made
during the annual naval appropriations debates between 1890 and 1914—the 52nd through the
63rd Congresses. Among other things, we coded each statement for the following two features:
(1) references to the international environment or to a particular state as threatening, or as having
actual or potentially hostile relations with the United States; (2) references to the international
environment or a particular state as unthreatening, or as having actual or potentially cooperative
relations with the United States. The frequency with which members provided these evaluations
should reveal something about their assessments of the international threat environment. Our
20
dependent variables in the analysis that follows are counts of the number of speeches in which
members gave one of these two kinds of evaluations.
A few typical examples help illustrate the types of statements included in our data. On
February 15, 1907, Representative John Weeks (R-MA) offered a negative evaluation of Japan,
The statement in question was as follows:
I am not for one moment considering the merits or demerits of this live question whether
we should or should not limit Japanese labor; but if we are not going to do it, if we are
not going to treat the Japanese as we treat the Hun and the Pole, and, in fact, every other
people of Europe, then we may expect that Japan will resent it. They may not resent it
with force now, but it is our duty to assume that they will later, and to prepare for just
that possibility (Congressional Record, 1907, Vol. 41, 3949).
By contrast, on March 26, 1910, Representative Isaac Sherwood (D-OH) gave the
following positive evaluation of the same country:
The spectacular jingoes are saying we must always be prepared for war, and they seem to
see a menace to peace in war-worn and war-tired Japan. This imaginary menace has
literally been worked to death. No intelligent American citizen who understands the aims
and mission of Japan has any fear of war. Even if Japan so desired, she would be utterly
unable to conduct a war with the United States (Congressional Record, 1910, Vol. 45,
3837).
Similarly, on February 15, 1907, Representative Theodore Burton (R-OH) gave the
following positive evaluation of the international environment, generally:
We have the respect of other nations. They look with admiration upon our growth,
surprising in its increase of power and prosperity beyond all that is recorded in the history
of progress. They depend upon us for supplies and many things essential both for peace
and war. No nation on the globe is threatening us; no nation on the globe will think
lightly of engaging in war with us (Congressional Record, 1907, Vol. 41, 3048).
In response to Representative Burton's remarks, Representative Adolph Meyer (D-LA)
provided his own negative evaluation of the international environment:
Unless human nature changes from all its tendencies during the ages, we can not [sic]
expect to realize [Mr. Burton's] ideals in the years to come. We may dream of universal
peace—we may repose calmly in the beatific hope that the millennium has been
reached—and that all the people of the earth have "beaten their swords into plowshares
21
and their spears into pruning hooks," but our awakening will tell us that wars on our
continent and with other nations have not ceased to exist (Congressional Record, 1907,
Vol. 41, 3055).
Overall, there were 5,199 speeches made during the naval appropriations debates
between 1890 and 1914. Because floor speeches were frequently interrupted, and members often
gave their remarks as part of on ongoing exchange with their colleagues, we treated all
statements made on the same day as part of the same speech. Roughly 30 percent of these
statements consisted of two sentences or less. Often these short statements were questions for
another member. Statements about international threat were relatively rare, occurring in 363 of
these speeches, roughly 7 percent of the total, or 12 percent of the speeches of more than two
sentences. The debates included a nearly identical number of positive and negative evaluations,
195 and 194, respectively. In 26 cases, members provided both a negative and a positive
evaluation in the same speech. General evaluations of the international environment were
slightly more common than specific assessments of other states. In some instances, members
offered both general and specific evaluations in the same speech. Comments about particular
states focused primarily on other major powers, though Spain received some attention before the
Spanish-American War, and Mexico received heightened attention during the final years of our
study, when the Mexican revolution got underway.
Third, we also consider a few individual-level effects in our analysis. We do not have
data on most of individual-level considerations discussed in research on threat perception. These
are not our focus in any event. However, we will include two background characteristics of
members of Congress that should influence their positions on the international threat
environment: college attendance and service as a military officer. Members who attended college
should be more aware of the international environment and more likely to take positions about it,
22
either positive or negative. Because of their professional training, and perhaps their desire to
justify greater military expenditures, military officers should be more likely to perceive
international threats than other members. These data are drawn from ICPSR and McKibben
(1997). Though these are not the only individual-level factors that ought to influence threat
assessment, they should provide a reasonable baseline against which to assess the size of the
effects associated with economic interests.
As we noted earlier, party and ideology might embody part of the effect of societal
interests. Societal interests might use these markers to select representatives likely to favor their
interests. If this selection process worked perfectly, these interests might have no need to
influence their representatives more directly by lobbying them once they were in office. The
distinction between selection and influence has some important implications for our research
design. First, rather than being a competing source of explanation, a member's party affiliation
might well reveal something about the influence of constituent economic interests. During the
1890–1914 period, party affiliation was a very good indicator of the positions a member of
Congress would take on many issues. As we noted earlier, Republican affiliation almost
perfectly predicted support for high tariffs. The leading figures associated with a more aggressive
and activist foreign policy during this period, including Henry Cabot Lodge, Theodore
Roosevelt, and others, tended to be Republicans. We are by no means the first to advance this
claim. Indeed, Bensel (2000) has extensively considered the relationship between the Republican
Party and the manufacturing sector. Nevertheless, testing this hypothesis for the period we will
consider is important because of the role party plays in transmitting the influence of the societal
economic interests. This leads to our first set of hypotheses.
23
H1. The size of the manufacturing sector in their home state is positively
associated with the probability of Republican victory in congressional elections.
H2. Republican Party affiliation should be positively associated with the number
of negative evaluations of the international environment.
H3. Republican Party affiliation should be negatively associated with the number
of positive evaluations of the international environment.
While selection is important, influence should also make a difference if the conflicting
interests of manufacturing and agriculture really affected the politics of foreign policy between
1890 and 1914. Though the Republican Party was closely associated with manufacturing
interests, Republicans did also represent some predominantly agricultural areas of the country.
Similarly, some Democrats were elected in areas where manufacturing predominated. This fact
gives societal economic interests ample reason to persuade their representative to move away
from the party's typical foreign policy position in many cases. In other words, even when we
control for party, we believe that economic interests in their home state should still be associated
with members' assessment of the international threat environment.
H4. The size of the manufacturing sector in their home state should be positively
associated with the number of negative evaluations of the international
environment.
H5. The size of the manufacturing sector in their home state should be negatively
associated with the number of positive evaluations of the international
environment.
Finally, we will compare our findings concerning party and economic structure to the
effects of two individual-level considerations that should influence threat perception during this
period. First, members of Congress who attended college should be more aware of the
international environment than were less educated members. We expect that they will be more
willing to provide both positive and negative evaluations of the international environment than
other members are. Second, members of Congress who had served as military officers should be
24
more attuned to international threats than members who lacked this experience. For example,
Richmond Hobson, whose grim assessment of the international environment we quoted at the
beginning of this paper, had a distinguished naval career before his election to Congress. We
chose to examine only former officers because members who had served in the ranks during the
Civil War—the primary source of overall military experience during this period—might have
had little cause to reflect on international threats.
H6. Members of Congress who had graduated from college should provide more
evaluations (positive and negative) of the international environment.
H7. Members of Congress who previously served as an officer in the military
should provide more negative evaluations of the international environment.
Empirical Analysis and Discussion
To test the hypotheses concerning the effect of economic structure on legislators' threat
assessments, we estimate two types of speech models. Using the legislator-congress as our unit
of analysis, we constructed two dependent variables counting the speeches that each legislator
gave during each congress containing evaluations of the level of threat in the international
environment. The dependent variable in our first speech model is the number of times each
legislator characterized the international environment in general or another state as hostile or
threatening. The second dependent variable is the number of times a legislator evaluated the
international environment or another state as benign or friendly. Using dependent variables that
capture both types of evaluations allows us to better assess how our variables of interest affect
attitudes toward the international environment. Because these dependent variables are counts and
likely to be overdispersed, we use a negative binomial model. In addition to our variables of
interest we also include fixed-effect dummy variables for each congress after the 52nd in order to
25
control for changes over time in the baseline number of speeches evaluating the international
environment.
Before testing our hypotheses about the effects of economic structure and personal
attributes on legislator attitudes, we first consider hypothesis 1 about the effect of economic
structure on the electoral fortunes of the two parties. Table 1 shows the results of a probit model
of Republican electoral success. 4 The Democratic Party dominated the heavily agrarian South
during this period as a legacy of the Civil War. To avoid the possibility that this pattern will
produce spurious support for our argument that manufacturing areas favored Republicans, we
include a dummy variable indicating whether the state was a former member of the Confederacy.
Finally, we also include fixed effects for each individual congress to control for the shifting
electoral fortunes of the two parties over time.
*** Table 1 about here ***
As the results in Table 1 indicate, the size of the manufacturing sector is significant and
positive. States with large manufacturing sectors were more likely to elect Republican
legislators. This relationship holds even though we control for Democratic dominance in the
South. However, this magnitude of this relationship is not overwhelming. For example, when
looking at a non-Southern district in the 56th Congress, increasing the manufacturing variable
from the 25th percentile to the 75th percentile (i.e. from .13 to .31) increases the predicted
probability of a Republican representative by 5 percentage points, from .65 to .70. Increasing the
4
Unit of observation is the legislator-election year. Note that we only included legislators who were elected through
a general election, or in some cases a special election. Legislators appointed by governors, for example, were
excluded. Also note that since the last election year preceding the observed congresses in our data is 1912, the
results from this model are driven primarily by the House, as very few senators were elected by direct election prior
to the ratification of 17th amendment.
26
manufacturing score from the 25th percentile (.13) to .50, the approximate maximum that we
observe in the data, increases the predicted probability of a Republican victory from .65 to .75.
These results suggest that party is unlikely to embody the entire effect of economic structure on
members' foreign policy outlooks. Selection mattered, but there is plenty of room for influence
effects of the sort we will consider next.
Having briefly assessed the impact of economic structure on the selection of legislators,
we now turn to our evaluation of the effect of economic structure and personal attributes on
assessments of the international environment. To begin, we will present some simple graphical
comparisons to illustrate the relationships between a state's economic structure and threat
assessments made by that state's legislators. The red and green lines in Figure 3 plot the average
number of evaluations legislators from each state provided per congress. The states shown in
Figure 3 are ordered according to the size of the manufacturing sector, with the smallest
manufacturing sectors on the left and the largest on the right. The shaded area indicates the
proportion of each state's workers employed in the manufacturing sector.
*** Figure 3 about here ***
Figure 3 provides some preliminary support for our basic expectations. Rhode Island, the
state with the largest manufacturing sector, has a rate of 0.5 negative evaluations per legislator
and no positive evaluations at all over the 24-year period that we examine. With the second
largest manufacturing sector, Massachusetts has a rate of approximately 0.9 negative evaluations,
and only about 0.4 positive evaluations per legislator. Maine and Delaware—the eighth and ninth
largest manufacturing states, respectively—both have approximately 1.4 negative evaluations per
27
legislator, but only 0.5 and 0 for positive evaluations. Washington, while not one of the most
manufacturing-intensive states, has the highest average rate of negative evaluations at 1.85, but 0
positive evaluations. Broadly speaking, the states with the highest rates of negative evaluations
are generally clustered on the right half of Figure 3. States on the left half more frequently have
few or no negative evaluations. Alternatively, Figure 3 also shows that positive evaluations are
distributed more evenly across the states listed. However, higher rates of positive evaluations
appear slightly more prevalent in the states that are less dependent upon manufacturing than in
those that are more dependent on it.
The bivariate comparisons are suggestive, but multivariate models permit more rigorous
hypothesis tests, as well as the comparison of economic structure with other possible influences
on threat assessment. Table 2 shows the results from two sets of models of the number of
negative and positive evaluations that legislators gave during each congress. Our primary interest
is the effect of economic structure, so models 1 and 2 evaluate the impact of the manufacturing
without additional variables for partisanship and individual characteristics. Models 3 and 4 then
introduce these controls for partisanship and individual characteristics, and serve as our base
models.
*** Table 2 about here ***
The results largely support our hypotheses about economic structure. The size of the
manufacturing sector performs as expected across all four of the models shown in Table 2. In
model 1, where we include only the manufacturing variable, a larger manufacturing sector is
positively correlated with the number of negative remarks a legislator made about the
28
international environment. This variable is also negatively correlated with the number of positive
remarks a legislator makes, as shown in model 2. Models 3 and 4 show that the effect of the
manufacturing sector is robust to the inclusion of variables for partisanship, college attendance,
and service as a military officer. These results support Hypotheses 4 and 5. As the qualitative
evidence reviewed earlier suggests, legislators representing states with large manufacturing
sectors tended to adopt a more hostile tone than those representing largely agricultural
constituencies. The preference of the manufacturing sector for the protection of domestic
markets and the acquisition of markets in less-developed areas implied that conflict with other
great powers was relatively likely. These states frequently sought preferential access to these
markets, potentially interfering with American access. By contrast, other great powers
constituted the principal market for American agricultural exports. Legislators representing
relatively agricultural states tended to view these states, and the international environment in
general, in a more benign way. 5
These results also provide some tentative support for the influence of partisanship on
threat assessment, as expressed in Hypotheses 2 and 3 above. Though we hypothesized that
Republicans should be more likely to view the international environment as threatening than
Democrats, the evidence does not support this claim. Party is not statistically significant in model
3. However, model 4 supports our hypothesis that Republicans should be less likely to give
5
Before the ratification of the 17th amendment in 1913, most Senators were not elected by popular vote. We have
considered that this might have implications for their relationship with their constituents, and have run several
robustness checks. Our findings are generally supported by these tests. Including a dummy variable for the Senate
does not alter our results in any way. Splitting the sample between the House and the Senate does yield some minor
changes in the Senate models. When predicting favorable evaluations the Republican Party variable remains highly
significant and negative, but manufacturing is not significant. Having attended college also appears to be negative
and significantly correlated with the number of positive evaluations given by a Senator. When predicting negative
evaluations manufacturing remains positive and highly significant in the Senate models. The Republican, college,
and officer variables all fail to attain statistical significance. In sum, our variable of primary theoretical interest—
manufacturing orientation—only fails to perform as expected when examining positive evaluations in the Senate.
Results from the House models are consistent with the pooled models.
29
positive evaluations of the international environment. The negative and significant coefficient
indicates that Republicans made fewer positive remarks about the international environment than
Democrats did, suggesting that they are less inclined to see opportunities for mutual gain and
cooperation. Alongside the evidence in Table 1 that the manufacturing sector was not a
substantively strong predictor of Republican electoral success, these results suggest that party
was not the primary vehicle for transmitting the foreign policy interests of the manufacturing
sector. To use Fordham and McKeown's terms, the effect of economic structure was primarily
about influence rather than selection.
The results from models 3 and 4 also indicate that the some of the personal attributes we
included had substantial effects on negative evaluations of the international environment. As we
expected in Hypothesis 7, having served as a military officer before entering Congress was
positively associated with the number of negative evaluations a legislator gave. On the other
hand, our expectations about college attendance were not supported. We hypothesized that
attending college should make members generally more aware of the international environment,
and thus more likely to take both types of positions. In fact, college attendance is not a
statistically significant predictor of either positive or negative evaluations.
*** Table 3 about here ***
We have also included four additional models in Table 3 as further robustness checks on
the primary models in Table 2. First, models 1 and 2 replicate our base models from Table 2, but
omit the Republican variable. Here we can see that when controlling for individual-level factors,
30
the results for manufacturing are consistent with those from Table 2. This result shows that the
effect of economic structure is robust to the inclusion or exclusion of party.
Second, though we are primarily interested in the effects of party and economic structure,
there are additional factors that might be correlated with evaluations of the international
environment. Models 3 and 4 include four additional control variables that might influence
legislators' evaluations. First, we add a control for whether or not the observed legislator is from
a coastal state. Given that our speech data are collected from naval appropriations debates,
legislators from coastal states might be more inclined to view the international environment as
threatening because of their greater exposure to potential naval threats. This issue was of
particular concern during the time period under consideration, because the Navy was not yet
large enough to deploy forces to both the Atlantic and the Pacific simultaneously. The Panama
Canal, which speeded up movement between the two oceans, was not completed until 1914.
Speeches by legislators from the Western coastal states, reveal the importance of this
consideration as the United States became more active in East Asia and as relations with Japan
worsened. Second, we include a dummy variable indicating whether or not the legislator was a
Senator. As members of a smaller body, Senators might have more likely to comment on
international affairs than members of the House. Third, we control for the size of the observed
legislator's state delegation to Congress. Because our data on manufacturing is only measured at
the state level, measurement error should be greater in larger states. The size of the legislative
delegation should capture the effects of this measurement error on our results.
Lastly, our focus on the influence of trade interests might seem to slight the effects of
legislators' economic interests in military spending itself. After all, this issue plays a major role
in analyses of the politics of national security policy in Congress during the postwar era (e.g.,
31
Mayer 1991; Rundquist and Carsey 2002). To address this possibility we also include a variable
indicating the share of employment in the iron and steel shipbuilding sector. While important in
theory, there is good reason to doubt its relevance in this particular historical setting. Military
spending before World War I was too small to have played the political role that some have
accorded it in more recent times. Balke and Gordon's (1989) estimates of the size of the U.S.
economy, coupled with historical data on the budgets for the War and Navy Departments (Carter,
et al. 2006, Table Ea636–643), suggest that overall military spending during the 1890–1914
period peaked at 1.65 percent in 1899, reflecting expenditures on the Spanish-American War. It
was less than one percent in most years. By contrast, military spending has not dropped below 4
percent of GDP since World War II.
The robustness checks in models 3 and 4 are consistent with our initial findings. The
results for our primary variables generally remain statistically significant and in the expected
direction. However, the manufacturing variable fails to attain statistical significance in model 4.
The size of the shipbuilding industry is not a statistically significant predictor of either positive
or negative evaluations. The variable indicating coastal states is positive and significant in model
3, though not in model 4. Legislators from coastal states were indeed more likely to view the
international environment as threatening, but controlling for this pattern does not substantively
change our results concerning economic structure. Finally, neither the Senate dummy variable
nor the size of the state's Congressional delegation is related to the number of speeches made in
either category.
It is worth noting that our base models, as well as our robustness checks, provide a
demanding test of the influence of economic structure. Our manufacturing variable is exogenous
to the other independent variables, but might have contributed to whether a former military
32
officer or Republican was elected. Including these other variables might thus cause us to
underestimate the effect of economic structure on legislators' evaluations. The effects estimated
in models 1 and 2 in Table 2, which exclude the other independent variables, are somewhat
larger. The fact that we still find manufacturing to be a significant predictor of evaluations in
most of the models that include other independent variables strengthens the case for
manufacturing's influence.
To better understand the impact that a state's economic structure has on legislators'
evaluations, the graphs in Figure 4 plot the predicted probability of a college educated
Republican legislator making at least one evaluation in a given congress. These predictions are
based on models 3 and 4 from Table 2. 6 The x-axis covers the approximate range of the
manufacturing variable observed in the data. In panel A of Figure 4, we see that legislators from
districts that are the least reliant on manufacturing have a probability of approximately 0.05 of
making at least one negative comment about the international environment. This figure increases
to 0.08 when we look at states where roughly 25 percent of the workforce is engaged in the
manufacturing sector. Legislators from the states most reliant on the manufacturing sector (50
percent) have a probability of 0.13 of making at least one negative remark. Further, the 95
percent confidence intervals shown around the predicted values not only indicate that the
predicted values are statistically significantly different from zero, but they also suggest that the
states that are the least reliant on manufacturing are statistically significantly different from the
states that are the most reliant on manufacturing.
Panel B in Figure 4 shows the predicted probability of observing at least one positive
evaluation of the international environment across the same range of manufacturing. Though the
manufacturing variable is negative and significant, its substantive effect for positive evaluations
6
All predicted values generated using CLARIFY (Tomz, Wittenberg, and King 2001).
33
is considerably smaller as compared to its effect on negative evaluations. It should also be noted
that Republicans were statistically significantly less likely than Democrats to offer positive
evaluations. There was no statistically significant difference between the parties for negative
evaluations, however. Legislators from states with relatively small manufacturing sectors had the
highest probability of making at least one positive evaluation, approximately 0.01. This
probability shrinks to about .006 at the high end of the manufacturing variable.
*** Figure 4 about here ***
Our results indicate that both economic structure and some personal attributes shaped the
attitudes of individual legislators during this time period. The relative size of these relationships
is informative. Figure 5 plots the predicted probability of observing at least one positive and one
negative evaluation. The base model shows the predicted probabilities for Democrats and
Republicans who were not college educated, had not served as military officers, and represented
a state with a manufacturing sector in the 25th percentile. Each subsequent category shows the
predicted probability of observing at least one negative or positive evaluation when we change
the values of the indicated variable. The dummies for college education and service as a military
officer change from 0 to 1. The manufacturing variable increases from the 25th percentile to the
75th percentile. Neither college education nor military service had a statistically significant effect
on the number of positive evaluations. As can be seen in panel A of Figure 5, party affiliation
had the largest substantive effect. The manufacturing sector is statistically significant and
negative, but, as both panel B in Figure 4 and panel A in Figure 5 suggest, its substantive effect
was relatively small.
34
*** Figure 5 about here ***
Panel B of Figure 5 provides a different picture. The baseline predicted probability of a
negative evaluation is roughly four to five times the baseline predicted probability of a positive
evaluation. While party mattered for positive evaluations, it had little effect on the probability of
a negative evaluation. For the sake of comparison we include college education in Figure 5,
which increased the predicted probability of observing at least one negative evaluation by
approximately 1.3 percentage points. The impact of a shift in manufacturing from the 25th to the
75th percentile has a slightly larger impact than having attended college. This change in
manufacturing increases the probability of at least one negative evaluation by approximately 2.5
percentage points. Finally, having served as a military officer prior to entering Congress led to an
increase over the base model of 4.7 percentage points. Of these variables, an individual's
experience as a military officer appears to have the greatest impact on the degree to which they
perceive a hostile and dangerous international environment. Of course, the manufacturing
variable could have a larger effect than that depicted in Figure 5, where we only look at the
affect of changes from the 25th to the 75th percentiles. This range represents a change between
two fairly common points in the data, but the impact of this variable could be much larger across
its full range, as Figure 4 indicates. Overall, these comparisons suggest that the impact of
economic structure on evaluations of the international environment typically fell somewhere
between that of college education and service as a military officer.
35
Conclusion
The evidence we have considered in this paper suggests that domestic economic interests indeed
influenced threat assessment. During the 1890-1914 period, representatives of the American
manufacturing sector, both in Congress and in groups like the National Association of
Manufacturers, sought both trade protection and greater access to overseas markets. These goals
prompted them to look primarily to markets in less developed areas of the world that would not
export manufactured goods to the United States. This foreign policy led its proponents to view
the other major powers with suspicion, both because American policy threatened the interests of
these states and because most of them were engaged in parallel efforts to gain privileged access
to less-developed markets on behalf of their own nationals. Legislators with ties to this sector of
the economy tended to offer negative evaluations of other major powers and of the international
environment in general. By contrast, representatives of more agricultural regions of the United
States saw these same great powers as their best customers. These interests led them to hold
more favorable views of the international environment, a pattern that is evident in the speeches
they gave during naval appropriations debates.
These economic interests were not the only influences on threat assessment, but they
were substantively important. We found that individual characteristics such as service as a
military officer also influenced legislators' evaluations of the international environment. While
we lack more refined data on the kinds of individual-level considerations cited in more recent
research on threat perception, we strongly suspect these things would also make a difference.
The estimated size of this particular individual effect was roughly comparable to that of
economic interests. In sum, both sets of considerations made a difference and deserve the
attention of future researchers.
36
Political party also mattered. Its effects were mostly but not entirely independent of
economic structure. Manufacturing districts were more likely to elect Republicans, but not
overwhelmingly so. The fact that Republicans were comparatively less likely to make positive
evaluations of the international environment is thus partly an indirect effect of economic
structure, but this selection process was not very powerful. In any event, the party effects on
threat assessment were not as large as we initially expected. Like the present, the 1890-1914
period was a time of relatively high party polarization. While these strong party divisions
influenced threat assessments, their effects were not as strong as one might have expected. Once
we controlled for the effects of economic structure, Republicans were no more likely to give
negative evaluations of the international environment than Democrats were. Moreover, even in
the case of positive evaluations, where party had a statistically significant and negative effect, its
impact was smaller than that of economic structure, college attendance, or military service.
The specific patterns discussed here pertain to a particular time and place. Should we
expect to find similar patterns in other historical settings? The theoretical logic behind the
influence of economic interests suggests that we should. Exposure to the international economy
generates winners and losers in other historical periods, and should thus help shape political
conflict over foreign policy. Nevertheless, at least two caveats are in order. First, the connection
between manufacturing sector, protectionism, and threat assessment was a function of the
structure of the international economy during the 1890-1914 period. This particular configuration
of economic interests is far from universal. The increasing competitiveness of American
manufacturing during and after World War I, and the emergence of political and military threats
to American access to European markets altered the foreign policy interests of those associated
with this sector of the economy. Second, the 1890-1914 period featured no global ideological
37
conflict comparable to that prevailing during the Cold War. Such conflicts might well exert a
very strong influence on the politics of foreign policy, but we cannot assess their importance
relative to economic interests here.
These cautionary notes aside, the role of economic interests deserves a place alongside
the individual-level considerations stressed in current research, especially when considering the
factors influencing threat assessment among policymaking elites. The importance of our findings
is highlighted by the fact that economic interests of the sort considered here unite groups of
people and are thus more likely to serve as the basis for political action than potentially
idiosyncratic characteristics that influence individual threat perception. By their very nature,
economic interests are linked to the resources needed to make their advocates politically
effective. The fact that members of Congress must consider the interests and opinions of their
constituents might attenuate the influence of their individual psychology and increase the
salience of economic interests in their home state when making policy. Accordingly, the broader
economic and societal factors that influence threat assessment in the setting we have considered
here might thus differ from what one would typically find in survey research on threat perception
in the general public. These factors are nevertheless very important because they have the
potential to directly affect the policymaking process.
38
References
Adams, Brooks. 1898. "The Spanish War and the Equilibrium of the World." Forum 25(6): 64151.
Adams, Brooks. 1900. America's Economic Supremacy. New York: The Macmillan Company.
American Sympathies in a Russo-Japanese War. 1904. Literary Digest 28(4): 101-2.
Atkinson, Edward. 1896. "The Cost of an Anglo-American War." Forum 21(1): 74-88.
Balke, Nathan S. and Robert J. Gordon. 1989. "The Estimation of Prewar Gross National
Product: Methodology and New Evidence." Journal of Political Economy 97(1): 38-92.
Bearce, David H. 2003. "Societal Preferences, Partisan Agents, and Monetary Policy Outcomes."
International Organization 57(2): 373-410.
Bensel, Richard Franklin. 2000. The Political Economy of American Industrialization, 18771900. New York: Cambridge University Press.
Berinsky, Adam J. 2009. In Time of War. Chicago: University of Chicago Press.
Blaine, James G. 1892. "The Presidential Election of 1892." North American Review 155(432):
513-25.
Burton, Theodore (OH). 1907. "Naval Appropriation Bill." Congressional Record Permanent
Digital Collection. Vol. 41. Page 3048. Accessed February 5, 2013.
Carter, Susan B., Scott Sigmund Gartner, Michael R. Haines, Alan L. Olmstead, Richard Sutch,
and Gavin Wright, eds. 2006. Historical Statistics of the United States. Millennial edition.
New York: Cambridge University Press.
Carter, Thomas H. 1894. "The Republican Outlook." North American Review 158(449): 423-32.
Cary, Clarence. 1898. "China's Complications and American Trade." Forum 25(1): 35-45.
Cohen, Raymond. 1978. "Threat Perception in International Crises." Political Science Quarterly.
93:1 93–107.
Conant, Charles A. 1900. "The United States as a World Power, I. The Nature of the Economic
and Political Problem." Forum 29(5): 608-22.
Coolidge, Archibald Cary. 1908. The United States as a World Power. New York: The
Macmillan Company.
Cutrone, Ellen A., and Benjamin O. Fordham. 2010. "Commerce and Imagination: The Sources
of Concern about International Human Rights in the United States Congress." International
Studies Quarterly 54(3): 633-55.
Eddy, Ulysses D. 1891. "Our Chance for Commercial Supremacy." Forum 11(4): 419-28.
Epstein, David, and Sharyn O'Halloran. 1996. "The Partisan Paradox and the U.S. Tariff, 18771934." International Organization 50(2): 301-324.
Farnham, Barbara. 2003. "The Theory of Democratic Peace and Threat Perception."
International Studies Quarterly. 47:3 395–415.
39
Fordham, Benjamin O. 2008. "Economic Interests and Congressional Voting on American
Foreign Policy." Journal of Conflict Resolution 52(5): 623-40.
Fordham, Benjamin O., and Katja B. Kleinberg. 2011. "International Trade and US Relations
with China." Foreign Policy Analysis. 7:3 217–236.
Fordham, Benjamin O., and Timothy J. McKeown. 2003. "Selection and Influence: Interest
Groups and Congressional Voting on Trade Policy." International Organization 57(3): 51949.
Goldstein, Judith. 1993. Ideas, Interests, and American Trade Policy. Ithaca: Cornell University
Press.
Gray, George. 1895. "Two Years of American Diplomacy." North American Review 160(461):
409-24.
Howell, William G., and Jon C. Pevehouse. 2007. While Dangers Gather. Princeton: Princeton
University Press.
Hoxie, R. F. 1903. "The American Colonial Policy and the Tariff." Journal of Political Economy
11(2): 198-219.
Inter-university Consortium for Political and Social Research and Carroll McKibbin. 1997.
Roster of United States Congressional Officeholders and Biographical Characteristics of
Members of the United States Congress, 1789-1996: Merged Data [Computer file]. 10th
ICPSR ed. Ann Arbor, MI: Inter-university Consortium for Political and Social Research
[producer and distributor], 1997.
Jones, Chester Lloyd. 1908. "The American Interpretation of the "Most Favored Nation"
Clause." Annals of the American Academy of Political and Social Science 32: 119-29.
Kehr, Eckart. 1977 [1927.] "Anglophobia and Weltpolitik." In Gordon Craig, ed., and Grete
Heinz, tr. Economic Interest, Militarism and Foreign Policy. Berkeley, CA: University of
California Press.
Kemmelmeier, Markus, and David G. Winter. 2000. "Putting Threat Into Perspective:
Experimental Studies on Perceptual Distortion in International Conflict." Personality and
Social Psychology Bulletin. 26:7 795–809.
Kleinberg, Katja B., and Benjamin O. Fordham. 2010. "Trade and Foreign Policy Attitudes."
Journal of Conflict Resolution. 54:5 687–714.
LaFeber, Walter. 1963. The New Empire. Ithaca, NY: Cornell University Press.
Latane, John Holladay. 1907. America as a World Power, 1897-1907. New York: Harper &
Brothers.
Lodge, Henry Cabot. 1895. "England, Venezuela, and the Monroe Doctrine." North American
Review 160(463): 651-8.
Low, A. Maurice. 1899. "Russia, England, and the United States." Forum 28(2): 172-81.
Low, A. Maurice. 1908. "Foreign Affairs." Forum 40(4): 307-14.
Maddison, Angus. 2003. The World Economy: Historical Statistics. Paris: Development Centre
of the Organisation for Economic Co-operation and Development.
40
Mahan, Alfred Thayer. 1987 [1890]. The Influence of Sea Power upon History, 1660-1783. New
York: Dover Publications.
Maoz, Zeev, and Bruce Russett. 1993. "Normative and Structural Causes of the Democratic
Peace." American Political Science Review. 87:3 624–638.
Mayer, Kenneth R. 1991. The Political Economy of Defense Contracting. New Haven: Yale
University Press.
McCormick, Thomas J. 1967. China Market. Chicago: Quadrangle Books.
McDonald, Patrick. 2009. The Invisible Hand of Peace: Capitalism, the War Machine, and
International Relations Theory. Cambridge: Cambridge University Press.
Menace of Japanese Success. 1905. Literary Digest 30(25): 920-1.
Meyer, Adolph (LA). 1907. "Naval Appropriation Bill." Congressional Record Permanent
Digital Collection. Vol. 41. Page 3055. Accessed February 5, 2013.
Moffett, Samuel E. 1899. "Ultimate World-Politics." Forum 27(6): 665-8.
Morgenthau, Hans J., revised by Kenneth W. Thompson. 1993. Politics Among Nations, brief
edition. Boston: McGraw-Hill.
National Association of Manufacturers. 1901. Proceedings of the National Reciprocity
Convention Held Under the Auspices of the National Association of Manufacturers of the
United States of America. Washington, D.C.: National Association of Manufacturers.
Our Duty to Save China from Japan. 1907. Literary Digest 35(9): 288.
Our Peaceful Contest with Japan. 1907. Literary Digest 35(17): 594.
Reid, W. C. Jameson. 1901. "The Political and Commercial Future of Asia." Forum 32(3): 34154.
Reid, Whitelaw. 1900. Problems of Expansion, as Considered in Papers and Addresses. New
York: The Century Company.
Reinsch, Paul Samuel. 1900. World Politics at the End of the Nineteenth Century. New York:
The Macmillan Company.
Republican National Committee. 1892. "Republican Party Platform." The American Presidency
Project. Accessed February 16, 2013. URL: http://www.presidency.ucsb.edu/platforms.php.
Republican National Committee. 1896. "Republican Party Platform." The American Presidency
Project. Accessed February 16, 2013. URL: http://www.presidency.ucsb.edu/platforms.php.
Republican National Committee. 1900. "Republican Party Platform." The American Presidency
Project. Accessed February 16, 2013. URL: http://www.presidency.ucsb.edu/platforms.php.
Republican National Committee. 1904. "Republican Party Platform." The American Presidency
Project. Accessed February 16, 2013. URL: http://www.presidency.ucsb.edu/platforms.php.
Republican National Committee. 1908. "Republican Party Platform." The American Presidency
Project. Accessed February 16, 2013. URL: http://www.presidency.ucsb.edu/platforms.php.
Republican National Committee. 1912. "Republican Party Platform." The American Presidency
Project. Accessed February 16, 2013. URL: http://www.presidency.ucsb.edu/platforms.php.
41
Rogowski, Ronald. 1989. Commerce and Coalitions. Princeton: Princeton University Press.
Rousseau, David L. 2002. "Motivations for Choice: The Salience of Relative Gains in
International Politics." Journal of Conflict Resolution. 46:3 394–426.
Rousseau, David L., and Rocio Garcia-Retamero. 2007. "Identity, Power, and Threat Percpetion:
A Cross-National Experimental Study." Journal of Conflict Resolution. 51:5 744–771.
Rundquist, Barry S., and Thomas M. Carsey. 2002. Congress and Defense Spending. Norman:
University of Oklahoma Press.
Russett, Bruce, and John Oneal. 2001. Triangulating Peace. W.W. Norton & Company.
Selbie, William. 1893. "Our Coming Rival." North American Review 157(441): 254-6.
Sherwood, Isaac (OH). 1910. "Naval Appropriation Bill." Congressional Record Permanent
Digital Collection. Vol. 45. Page 3837. Accessed February 1, 2013.
Taussig, F. W. 1910. The Tariff History of the United States. 5th ed. New York: G. P. Putnam's
Sons.
Tomz, Michael, Jason Wittenberg, and Gary King. 2001. CLARIFY: Software for interpreting
and presenting statistical results. Version 2.0 Harvard University.
Trask, David F. 1981. A Short History of the U.S. Department of State, 1781-1981. Washington,
D.C: U.S. Dept. of State, Bureau of Public Affairs, Office of Public Communications.
U.S. National Archives and Records Administration. 2011. "U. S. Electoral College." Available
at: http://www.archives.gov/federal-register/electoral-college/historical.html [Accessed
February 13, 2011].
United States Census Bureau. 1901. "1900 Census of Population and Housing." 1900 Census of
Population and Housing. Available at:
http://www.census.gov/prod/www/abs/decennial/1900.html [Accessed February 25, 2011].
Viner, Jacob. 1924. "The Most-Favored-Nation Clause in American Commercial Treaties."
Journal of Political Economy 32(1): 101-29.
Walt, Stephen M. 1987. The Origins of Alliances. Ithaca, NY: Cornell University Press.
Weeks, John (MA). 1907. "Naval Appropriation Bill." Congressional Record Permanent Digital
Collection. Vol. 41. Page 3949. Accessed February 1, 2013.
Williams, William Appleman. 2009 [1959]. The Tragedy of American Diplomacy. 50th ed. New
York: W.W. Norton & Co.
42
Table 1
Probit Model Predicting Election of Republican
Representative
Manufacturing
0.724***
(0.230)
South
-1.982***
(0.0788)
Constant
-0.554***
(0.104)
Observations
4695
Robust standard errors in parentheses. Fixed
effects for individual congresses not shown.
*** p<0.01, ** p<0.05, * p<0.1
43
Table 2
Negative Binomial Models Predicting Legislator Evaluations of the International Environment
Manufacturing Only
Base Model
(1)
(2)
(3)
(4)
Negative
Positive
Negative
Positive
Evaluations
Evaluations
Evaluations
Evaluations
Manufacturing
2.265***
-2.155***
2.608***
-1.251*
(0.670)
(0.733)
(0.663)
(0.735)
Republican
-0.115
-0.548***
(0.174)
(0.186)
Attended College
0.277
0.0251
(0.176)
(0.181)
Officer
0.824***
0.0554
(0.198)
(0.219)
-2.560***
-4.600***
-2.643***
Constant
-4.050***
(0.348)
(0.246)
(0.347)
(0.281)
***
***
***
1.446
1.901
1.338
1.860***
ln 
(0.284)
(0.232)
(0.266)
(0.237)
Observations
5846
5846
5846
5846
Robust standard errors in parentheses. Fixed effects for individuals congresses not shown.
*
p < 0.10, ** p < 0.05, *** p < 0.01
44
Table 3
Negative Binomial Models Predicting Legislator Evaluations of the International Environment
Base Model Excluding
Base Model with
Republican Variable
Controls for State
Characteristics
(1)
(2)
(3)
(4)
Negative
Positive
Negative
Positive
Evaluatio Evaluatio Evaluatio Evaluatio
ns
ns
ns
ns
***
***
***
Manufacturing
2.439
-2.142
2.127
-0.925
(0.643)
(0.731)
(0.753)
(0.878)
Republican
0.0545
-0.565***
(0.189)
(0.198)
Attended College
0.278
0.0253
0.276
0.00942
(0.176)
(0.182)
(0.177)
(0.176)
***
***
Officer
0.820
0.0256
0.758
0.0109
(0.198)
(0.217)
(0.196)
(0.225)
Shipbuilding
-92.51
-143.0
(91.11)
(150.5)
Coastal State
0.591***
-0.0164
(0.179)
(0.171)
Senate Dummy
0.119
0.209
(0.190)
(0.196)
Size of Delegation
-0.00130
0.00545
(0.00736) (0.00930)
Constant
-4.600***
-2.587***
-4.842***
-2.758***
(0.347)
(0.284)
(0.338)
(0.305)
1.341***
1.899***
1.281***
1.853***
ln 
(0.267)
(0.233)
(0.252)
(0.243)
Observations
5846
5846
5846
5846
Robust standard errors in parentheses. Fixed effects for individual congresses not shown.
*
p < 0.10, ** p < 0.05, *** p < 0.01
45
Figure 1.
Proportion of Workforce in the Manufacturing Sector, 1900
0.26
WA
0.42
NH
0.31
MT
0.08
ND
0.18
MN
0.21
OR
0.22
ID
0.24
UT
0.30
AZ
0.40
PA
0.16
IA
0.27
IL
0.32
CO
0.14
KS
0.07
OK
0.15
NM
0.00 to 0.13
0.13 to 0.22
0.22 to 0.30
0.30 to 0.54
0.30
OH
0.23
IN
0.19
MO
0.17
VA
0.13
NC
0.02
TN
0.07
AR
0.10
SC
0.10
AL
0.46
CT
0.40
NJ
DC
0.21
WV
0.14
KY
0.05
MS
0.08
TX
0.47
MA
0.25
MI
0.13
NE
0.25
NV
0.32
ME
0.35
NY
0.24
WI
0.10
SD
0.22
WY
0.25
CA
0.27
VT
0.10
GA
0.10
LA
0.16
FL
0.31
DE
0.29
MD
0.53
RI
Figure 2.
U.S. Exports by Selected Commodities
40%
Cotton
Percent of Total Exports
35%
Automobiles, iron and steel, and
machinery
30%
25%
20%
15%
10%
5%
0%
1875
1880
1885
1890
1895
1900
1905
Source: Historical Statistics of the United States
1910
1915
1.6
Tennessee
Mississippi
Oklahoma
Arkansas
Texas
Alabama
North Dakota
South Dakota
New Mexico
Georgia
South Carolina
Louisiana
Idaho
Nevada
Kansas
North Carolina
Wyoming
Kentucky
Nebraska
West Virginia
Iowa
Florida
Virginia
Arizona
Utah
Montana
Minnesota
Missouri
Oregon
Colorado
Washington
California
Indiana
Michigan
Wisconsin
Illinois
Vermont
Ohio
Maryland
Delaware
Maine
Pennsylvania
New York
New Jersey
New Hampshire
Connecticut
Massachusetts
Rhode Island
Evaluation Rate
2
1.8
Positive Evaluations
Negative Evaluations
0
0.8
1.4
0.7
1.2
0.6
1
0.5
0.8
0.4
0.6
0.3
0.4
0.2
0.2
0.1
0
Manufacturing Sector Size
Figure 3.
Positive and Negative Evaluations and Manufacturing Sector Size
1
0.9
Figure 4 – Predicted Probability of Non-Zero Counts by Manufacturing Sector Size
50
Figure 5 – Predicted Probability of Non-Zero Counts by Individual Attribute.
51