Resident Networks and Corporate Connections: Evidence from World War II Internment Camps

• Published date: 01 February 2017
• Date: 01 February 2017
• Author: CHRISTOPHER MALLOY,LAUREN COHEN,UMIT G. GURUN
• DOI: http://doi.org/10.1111/jofi.12407

THE JOURNAL OF FINANCE •VOL. LXXII, NO. 1 •FEBRUARY 2017

Resident Networks and Corporate Connections:

Evidence from World War II Internment Camps

LAUREN COHEN, UMIT G. GURUN, and CHRISTOPHER MALLOY∗

ABSTRACT

Using customs and port authority data, we show that firms are significantly more

likely to trade with countries that have a large resident population near their firm

headquarters, and that these connected trades are their most valuable international

trades. Using the formation of World War II Japanese internment camps to isolate

exogenous shocks to local ethnic populations, we identify a causal link between local

networks and firm trade. Firms are also more likely to acquire target firms, and report

increased segment sales, in connected countries. Our results point to a surprisingly

large role of immigrants as economic conduits for firms.

FIRMS BUY AND SELL GOODS in a global marketplace. Nearly half of all S&P

500 firms’ sales, for instance, come from abroad. Understanding how firms

differentially navigate this marketplace is critical to identifying which firms

will ultimately succeed, and hence how investors should allocate capital among

these firms. Success in a global setting depends not only on the goods or services

that firms provide, but also on the various networks that firms can exploit to

access foreign markets.

In this paper we investigate one type of network that firms can employ in

accessing foreign markets, namely, local resident networks. Specifically, we ex-

ploit variation in ethnic population breakdowns across metropolitan statistical

areas (MSAs) in the U.S. to investigate whether local residents’ ties to their

home countries play a role in creating important bilateral country linkages for

firms headquartered in these areas. We show that local resident networks have

a first-order impact on each of the primary ways in which corporations operate

∗Cohen is at the Harvard Business School and NBER, Gurun is at the University of Texas at

Dallas and NBER, and Malloy is at the Harvard Business School and NBER. We thank Joshua

Aizenman, Gennaro Bernile, Geoffrey Booth, Lee Branstetter, Bill Cready, Robert Feenstra, Fritz

Foley, Rob Hansen,George Korniotis, John McLaren, Florian Peters, Ken Singleton, Youngxiang

Wang,two anonymous referees, and seminar participants at the 2014 AFA meeting in Philadelphia,

the 2013 NBER International Trade and Investment Meeting at Stanford, the SFS Cavalcade at

University of Miami, Brigham YoungUniversity, Australian National University, Ozyegin Univer-

sity,Michigan State University, Sabanci University,University of Alabama, University of Arizona,

Baruch College, University of Amsterdam/Duisenberg School of Finance, University of Mannheim,

University of Melbourne, University of Miami, University of South Florida, University of Texas

at Dallas, University of Virginia, Washington University in Saint Louis, and Tulane University

for comments. We are grateful for funding from the National Science Foundation. This paper also

circulated under a previous title: “Resident Networks and Firm Trade”. We have read the Journal

of Finance’s disclosure policy and have no conflicts of interest to disclose.

DOI: 10.1111/jofi.12407

207

208 The Journal of Finance R

globally: import and export behavior, international M&A activity, and segment

sales abroad, as well as on the value implications of such trade decisions.

We begin by using micro-level import and export data collected from customs

and port authorities. Combined with census-tract demographic data, these

micro-level trade data allow us to link individual firms’ trade decisions to their

local resident populations. For the universe of publicly traded U.S. firms over

our 20-year sample period, we find that firms export more to—and import

more from—countries with which they have stronger local resident network

connections.

Because residents’ location decisions are themselves influenced by firm-level

trade activity and the factors that drive this trade activity, the effect of resi-

dent location is endogenously related to the factors that cause these locations

to change. To address this endogeneity concern, we identify a set of plausibly

exogenous shocks to population residence—the forced relocation of Japanese

and Japanese-Americans into Japanese internment camps during World War

II—and show that these shocks have a large and significant impact on firm-

level trade decisions. The Japanese internment camps, established throughout

the country to house Japanese and Japanese-Americans from the West Coast

following the bombing of Pearl Harbor in 1941, represented a sizable shock to

the Japanese populations surrounding them. For example, in the 1940 census

(pre-internment camps), the Japanese population of Arkansas was three peo-

ple. Arkansas was later a site of two internment camps—a shock of over 17,000

Japanese residents into the state. Such shocks have had an enduring impact

on the affected areas as many internees ultimately settled around these camps,

having no home or work to return to after the war ended. Indeed, MSAs sur-

rounding World War II Japanese internment camps have roughly three times

(t=9.19) higher Japanese populations today than MSAs that did not house

internment camps. In addition, we show that the MSAs surrounding World

War II Japanese internment camps have over three times (t=4.34) as many

sister cities to Japan as similar cities throughout the rest of the U.S. Further,

as a placebo test, we examine the growth of other Asian ethnicities in the

same locations surrounding internment camps. We find no evidence that they

grew, nor is there any significant Asian population (besides Japanese) in these

surrounding areas today. Taken as a whole, our evidence suggests that the

Japanese internment camps of World War II represent significant exogenous

shocks to Japanese populations that persist to the present day.

Having established their exogeneity, we examine the effect of new Japanese

populations on firms’ trade decisions. Firms in MSAs surrounding internment

camps import and export significantly more to Japan today than other firms. In

terms of magnitude, a one-standard-deviation increase in exogenous Japanese

population increases exports by 67% (t=2.81) and imports by 101% (t=4.51).

We next show that our findings extend across the entire universe of U.S. firms

and firm-country trade destinations over a nearly 20-year sample period. To do

so, we exploit novel import and export data collected through public records that

must be reported by shippers and are subsequently made publicly available by

customs and port authorities. We use these data to ask whether immigration

Resident Networks and Corporate Connections 209

patterns that result in concentrated ethnic populations close to certain firms

give rise to strategic trade decisions for these firms. We measure firm-country

networks as the share of residents in the MSA in which a firm’s headquarters

is located (hereafter, a firm’s MSA) that have the same ethnicity as the country

to/from which the firm is exporting/importing (a variable we call Connected

Population). We find evidence that firms export more to and import more from

countries with which they have stronger information links. Specifically, we

show that a one-standard-deviation increase in Connected Population increases

the amount the firm exports to (imports from) a country by 63%, t=4.93 (60%,

t=5.59).

The increased exports (and imports) associated with resident networks leads

to increased sales and profitability. For instance, when we compare the prof-

itability of “strategic exporters,” that is, firms that export to a country with

which they have a large connected population, and “nonstrategic exporters,”

firms that export the same amount to the same country but do not have a large

connected population, we find that strategic exporters experience a statistically

significant 11% increase in their future profitability (EBITDA/Assets) relative

to nonstrategic exporters.

We next show that the effect of local resident networks in international

transactions is not confined to imports and exports. Local resident networks

also have large and significant effects on M&A activity and segment sales

in connected countries. For example, firms are significantly more likely to

purchase target firms in countries they are connected to through their local

resident networks. Moreover,using information disclosed in segment filings, we

show that firms are more likely to have an international presence in countries

that they are connected to through their resident networks.

We further explore how information is transferred across resident networks.

While we cannot obtain the ethnic makeup of the firms’ entire employee base,

we are able to collect the ethnic makeup of sample firms’ board of directors

(including top management). These data allow us to identify a specific channel

through which resident populations can influence firm decisions—connected

boards of directors. We find that local ethnic population is a strong predictor

of a board’s ethnic makeup (e.g., if there is a larger Chinese population in a

given firm’s MSA, the firm’s board is significantly more likely to have Chinese

board members). More importantly for our purposes, we show that when a

strategic importer (exporter) has a connected board member on its board, it

trades significantly more with the connected country.For instance, firms export

68% more than the median firm (t=4.03) to countries that they are connected

to with a connected board member.

Finally, to shed further light on the mechanism behind our findings, we

test whether strategic trading using local resident networks is more pro-

nounced when these networks are more valuable. To do so, we first exam-

ine the role of tariff controls between the U.S. and a given connected coun-

try for a given product. Consistent with lower (higher) tariffs increasing

(decreasing) the value of local resident networks, we find significantly more

strategic trading (i.e., imports from connected countries) when U.S. import