Voyages of Discovery and the Rise of Piracy
in Sixteenth-Century China
James Kai-sing Kung and Chicheng Ma∗
This version, August 2012
ABSTRACT
The Voyages of Discovery of the late 1490s led to unprecedented trade expansion
in Europe but gave rise to the curious ascent of piracy in China. By analyzing a
uniquely constructed dataset, we show that while the coastal prefectures with
greater trade potential were 1.5 times more likely to be subjected to large-scale
plundering than those less likely to engage in trade, this was only true when the
imperial authorities of the Ming dynasty committed to attaining economic autarky by enforcing the “sea ban” policy most rigorously (1550-1567), but not before they did so or after the policy’s abolishment.
∗
We thank Ying Bai, Philip Hoffman, Debin Ma, Nathan Nunn, Jeffery Williamson, and seminar
participants at various universities for helpful comments. We alone are responsible for any remaining
errors. James Kai-sing Kung, Division of Social Science, Hong Kong University of Science and Technology,
Clear Water Bay, Hong Kong. Email: [email protected]. Chicheng Ma, School of Economics, Shandong
University, 27 Shanda Nanlu, Jinan, Shandong Province, China. Email: [email protected].
1. Introduction
Piracy typically entails plundering ships at sea. This practice gained global prominence
in the sixteenth century. By the seventeenth century, piracy had grown from mere sporadic
conflicts in the seas of Mediterranean and Scandinavia to frequent occurrences on nearly all
major trade routes connecting the world’s continents (Gosse, 1932; Andrews, 1964; Leeson,
2007).1 Curiously, the world’s largest wave of piratical attacks occurred in Chinese waters
around the mid-sixteenth century, at a time when Europe experienced an unprecedented
expansion in overseas trade. These pirates raided the coastal cities and towns of China for
silk and other popular exports and shipped their plunder to foreign markets in Macau, Japan, and Southeast Asia for trade (So, 1975).2 So powerful were the Chinese pirates that they
defeated the imperial troops and took control of the southern seaboard where the bulk of
maritime trade occurred (Anderson, 1995), allegedly ruining many prosperous coastal
towns in the process (Von Glahn, 1996; Mann, 2011).3
While Europe’s trade expansion that followed the opening of the Atlantic trade routes
(circa the 1490s) had likely sparked a return to piracy (e.g., Braudel, 1992; Anderson, 1995;
Cordingly, 1996; Starkey, 2001; Pomeranz and Topik, 2006), what is puzzling is that incidence was disproportionately high in the territorial waters of an altogether different region—China. We examine this puzzle by employing a uniquely constructed data set. We
show that, while Europe’s trade expansion into Asia increased the potential returns to raid,
it was only a necessary condition behind the sharp rise in maritime conflicts over mercantilist claims in the Chinese context. By restricting (maritime) trade, China’s “sea ban” (haijin)
policy effectively diverted what originally was legitimate, mercantile interest to piratical activities—the latter conducted in attempt to capture the otherwise dissipated economic rents.4
A unique advantage of examining Europe’s overseas trade expansion into China, as was
evident from the massive influx of European merchant-explorers along the China coast after
1
After 1500, pirates flourished in the Atlantic, the Caribbean, the Indian Ocean, and Southeast Asia (Gosse,
1932).
2
As these illegal activities in China occurred on the land by people who operated from boats, we refer to
them as piratical attacks.
3
While the Chinese pirates that wreaked havoc during the mid-sixteenth century may not be as
(in)famous as their Atlantic and Caribbean counterparts in 1650-1730—a period during which the West
experienced the largest wave of piracy thus far, they were by far the bigger group as at their peak they
had 100,000 crew members whereas their Atlantic and Caribbean counterparts had a mere 5,000 (Chao,
2005; Travers, 2007). According to some historians the magnitude of piratical activities in China then—be
it measured in terms of crew size or organized capacity—has remained unsurpassed to this day (Antony,
2003).
4
Xu Fuyuan, governor of Fujian Province in the 1580s, once said that “when the (foreign trade) market is
open pirates become merchants, and when the market is closed merchants become pirates”, i.e. pirates
and merchants were basically the same people according to him (Chao, 2005, p. 203).
1
1517 in search for a wide range of Chinese commodities,5 is that to the Chinese it was effectively an exogenous shock. This “demand shock” generated by the Europeans thus allows
us to identify an exogenous impact of Europe’s overseas trade expansion on the rise of piracy
in China, a study rendered possible by the completeness of official records kept by the Ming
authorities in regard to the year and place of each piratical attack. This information enables
us to construct and analyze a unique panel data set covering all 33 coastal prefectures for the
period 1371-1640.
Our strategy is to adopt essentially a difference-in-differences approach for empirical
analysis. We examine whether prefectures with greater trade (raid) potential actually suffered from greater plundering by the pirates after 1500 as a result of the Europeans’ sudden
demand for Chinese goods. In particular, in the light of the immense popularity of silk, we
use “whether a coastal prefecture in China was a major silk-producing area (silk center)” as
our primary measure of a prefecture’s trade (raid) potential. Additionally, in order to fully
capture a prefecture’s trade (raid) potential, we employ two alternative measures: the first is
“whether a prefecture had already developed a port for conducting foreign trade prior to
the Ming dynasty”, whereas the second pertains to “a prefecture’s prevailing urbanization
rate”. The underlying assumption for choosing these alternative proxies is that they were
highly correlated with the level of exports (Skinner, 1977; Li, 2000; Xu and Wu, 2000), making the prefectures concerned more likely targets of piratical raids.
We control for a number of covariates that may be correlated with the incidence of piracy. First, piracy could have been caused by economic hardships rather than by the interplay
of trade shocks and China’s autarkic policy. To rule out this possibility we control for it using the number of famines as the pertinent proxy. Likewise, piracy may also be a response
to growing population pressure, hence it is necessary to control for a prefecture’s changing
population density over time. Last but not least, to the extent that coastal defense acts as a
major deterrent to piracy, we employ the number of naval fortresses stationed on the coastline to proxy for the variation in defense capacity over time.
Our empirical results thus obtained clearly demonstrate that the rise in piracy in midsixteenth-century China was not the sole outcome of Europe’s trade expansion, but was due
to its interplay with China’s policy of rigorously suppressing maritime trade. This is evidenced by the fact that prefectures with greater trade (raid) potential did not actually suffer
from greater plundering by the pirates until 1550—the year when China began to enforce its
“sea ban” policy in earnest, thanks to the initially favorable response of the Chinese merchants to engage in trade with the Europeans on the one hand, and the initial tolerance of
the Ming authorities on the other.6 As maritime trade—especially with the Portuguese—
5
Silk was the most popular item, followed by tea and porcelain (Wills, 1993; Brook, 1998; Ma, 1998).
The expansion of European trade around 1517 (especially with the massive arrival of the Portuguese)
had resulted in the rapid development of what originally were sleepy island towns (Shuangyu and
6
2
expanded, Emperor Jiajing became increasingly concerned that the imperial “sea ban” policy
could no longer be upheld, at which point (around 1550) he decided to crack down on what
he considered “illicit trade”.7 The result was, up to 1,200 “illicit” boats were scuttled and
hundreds of “smugglers” killed (Wills, 1979; Brook, 1998). Moreover, a compulsory neighborhood mutual-monitoring scheme known as pai jia was introduced to the coastal provinces, under which those who were caught engaged in maritime trade would not only face the
death sentence themselves, they also jeopardized the lives of their family members and
neighbors (Hu, 1987). But even so, the potential profits of plundering the restricted goods
only sent the merchants into piracy. Indeed, our estimations find that piratical attacks on the
silk centers after 1550 were 1.5 times more likely than piratical attacks on the non-silk centers (year on year for each prefecture). Specifically, Europe’s trade expansion—proxied by
the tonnage of European commercial ships that arrived in Asia—explained more than half
of the difference in the number of piratical attacks sustained (of 1.5 times) between the silk
centers and the non-silk centers during 1551-1567.
The importance of the “sea ban” in contributing to the phenomenal rise in privacy is
clearly borne out by the evidence that piracy plummeted drastically after 1567 as soon as the
imperial authorities legalized private foreign trade in that year. From then onwards until the
end of the Ming dynasty, piracy returned to the same low levels as it had exhibited for nearly two centuries prior to 1550. From 1567 to 1640 Chinese maritime commerce flourished
(Xu and Wu, 2000; Wang, 2007).8
However, our key explanatory variable—a prefecture’s trade (raid) potential—is endogenous. A main source of endogeneity is that our estimations may have omitted some variables that may be correlated with both piracy and trade (raid) potential. To address this
concern, we exploit the exogenous variation in trade potential among China’s coastal prefectures as our instrumental variable. Premised on the fact that mulberry leaves were a critical
input in the upstream production process of sericulture in the historical context of China, we
employ the share of land with loamy soil—the texture of which is most suitable for cultivating mulberry trees—to proxy for the suitability of planting this crop to instrument the silk
center variable. The instrumented results remain robust.
Another concern is that our policy variable of “sea ban” might be endogenous, to the extent that trade was legalized in 1567 in response to the heightened piratical activities (von
Yuegang) off the coasts of Southeast China into prosperous metropolises (Figure A1; see also Wills, 1993;
von Glahn, 1996; Brook, 1998).
7
China implemented the said “sea ban” policy in 1368, which effectively prohibited the Chinese people
from engaging in foreign (maritime) trade. The alleged purpose of the “sea ban” was to avoid the
potential coastal unrests triggered possibly by the interaction between the Chinese and the foreigners
(Fairbank, 1968; Chao, 2005).
8
Maritime trade was once again prohibited in the 1650s with the succession of the Qing dynasty (16441911). From then onwards China remained closed to the outside world until the outbreak of the first
Opium War (1839-42). Since systematic data on piracy in the Qing period are not available, we are unable
to examine the effect of the autarkic policy under the Qing rule on the incidence of piracy.
3
Glahn, 1996). To address this concern, we lag the number of piratical attacks. Moreover, to
ensure that the periods selected for our difference-in-differences analysis are not contaminated by the “spillovers” of the Japanese political turmoil during 1467-1590 (the Warring
States period when the lack of centralized political control may have predisposed warlords
to prey along the China coast), we interact trade potential with the post-1467 and post-1590
time dummies as additional controls. The results after controlling for all these additional
covariates remain robust.
In addition to contributing to the small literature that examines piracy in the context of
trade (e.g., Hillmann and Gathmann, 2011) or in connection to political regimes (e.g., Iyigun
and Ratisukpimol, 2011),9 our study also contributes to the growing literature that examines
the economic impact of Atlantic trade or more generally the “Columbian Exchange” on the
overseas trade expansion of Western Europe after 1500 (O’Rourke and Williamson, 2002), on
urbanization and population growth in the “Old World” (Acemoglu, Johnson and Robinson,
2005; Nunn and Qian, 2011), and on the changing welfare and living standard in Europe
(Hersh and Voth, 2009), among others.10 Our study examines, we believe for the first time,
the impact of the new trade routes on the rise and fall of maritime conflicts in a China that
was eager to attain autarky.
The remainder of our paper proceeds as follows. Section 2 provides the historical background on Europe’s overseas trade expansion after 1500 and China’s “sea ban” policy implemented during the Ming dynasty, and develops on these bases the hypothesized reasoning behind the rise and fall of piracy along China’s coasts. Section 3 introduces the data
sources and describes the variables of interest. The empirical strategy employed and the
main empirical results are then discussed in Section 4, whereas Section 5 reports both the
instrumented results and those obtained with the inclusion of a richer set of controls. Section
6 concludes.
2. Historical Background
Thanks to Vasco da Gama’s discovery of the new sea route rounding the Cape of Good
Hope to Asia and Christopher Columbus’ discovery of the New World in the late 1490s, the
world’s continents became directly connected for the first time in human history. From 1500
onwards, Western European nations—mainly the Atlantic trading powers comprising Britain, France, the Netherlands, Portugal, and Spain—became actively involved in overseas
9
For instance, Hillmann and Gathmann (2011) find that the declining profitability of raiding in seventeenth- and eighteenth-centuries Britain had led to the decline of such activities, whereas Iyigun and
Ratisukpimol (2011) find that piracy is most likely to occur in countries ruled by authoritarian regimes, as
such countries typically have less liberal trade regimes.
10
Nunn and Qian (2010) provide a review of this emerging literature.
4
trade with Asia and the New World (Frank, 1998; O’Rourke and Williamson, 2002). Using
the total tonnage of European ships that arrived in Asia and the Americas to proxy for the
volume of trade, Figure 1 shows that European foreign trade volume indeed grew rapidly
after 1490. The total tonnage of European ships that arrived in Asia and the Americas increased sharply from virtually non-existent before 1500 to 63,893 tons during 1500-1510, and
kept growing ever since at the rate of 10.2% per decade (de Vries, 2003).11
[Figure 1 about here]
The discovery of new sea routes rounding the Cape of Good Hope to Asia in the late
1490s enabled the Europeans to expand their trade into China, whose products—
particularly silk, porcelain and tea— enjoyed immense demand in Europe at the time (Elvin,
1973; Wills, 1993; Brook, 1998, Ma, 1998). This is evidenced by the massive influx of European merchant-explorers to the China coast during 1517 and 1550.12 Some islands off the
coastal provinces of Fujian and Zhejiang (Figure A1), for instance, which prior to 1517 were
nothing more than sleepy towns, became bustling metropolises by the 1540s (Wills, 1993;
Brook, 1998).13
Unfortunately, China’s expanding foreign trade took place within the context of a “sea
ban” policy implemented upon the founding of the Ming dynasty in 1368. This policy effectively prohibited the Chinese people from engaging in foreign or specifically maritime trade.
The so-called trade in Ming China was confined to primarily “tributary trade”—a form of
limited “commerce” in which China asserted itself as the hegemonic power, with the vassal
states making periodic trips of homage to the imperial court to offer gifts to the emperor in
exchange for a limited amount of goods (Fairbank, 1968). There are different views regarding the imposition of the “sea ban”. One view attributes the ban to the emperors’ concern for
potential coastal unrests triggered possibly by interactions between the Chinese and the foreigners (Fairbank, 1968; Chao, 2005); the other view sees the ban as reflecting the emperors’
inherent lack of interest in developing overseas trade in light of China’s autarkic economy
11
Among the Europeans, the Portuguese were the earliest to arrive in Asia. After seizing Goa of India in
1510 and Malacca in 1511 (Figure A1), the Portuguese gradually controlled trade between Asia and
Europe. Then came the Spanish, who, after occupying Manila of the Philippines in 1571 established the
Pacific trading route linking Asia to the Americas (the famous Manila Galleon trade). The Dutch, the
British, and the French were the last to arrive in Asia (from the late sixteenth century onwards).
12
Approximately 180 Europeans (mainly Portuguese) were stationed on the islands off the Chinese coast
in the 1520s, but by the 1550s the number had increased to over 600, and to about 1,000 by the 1620s. These
figures are taken from the rough estimates of Chao (2005), Ljungstedt (1832), and Ptak (1982).
13
The lucrative profitability of trading in Asia in general and with China in particular is well illustrated by
the evidence that, whereas 1 dan (approximately 50 kg) of raw silk was sold for 100 silver liang
(approximately US$20 today) in China’s lower Yangzi region, it could be sold for 500 silver liang in the
Philippines in the 1580s (Quan, 1986). See also Findlay and O’Rourke (2007).
5
and the underlying Confucian ideology of “putting agriculture before business (zhongnong
qingshang)” (Brook, 1998; Landes, 1998).14
Initially, the imperial authorities tolerated the budding trade resulting from the Voyages
of Discovery. But by around the 1550s, the authorities had had enough (von Glahn, 1996). To
demonstrate their determination to suppress the growing maritime trade, the Ming armies
demolished what they saw as smuggling activities along the China coast in 1548 and 1549,
scuttling more than 1,200 “illicit” boats and killing hundreds of “smugglers” (Wills, 1979;
Brook, 1998). At the same time, to prevent further smuggling activities from occurring, the
imperial authorities imposed on the coastal provinces a neighborhood mutual-monitoring
scheme known as pai jia, under which every ten households were organized into a single
unit of pai for the sake of mutually monitoring and preventing each other from secretly engaging in foreign trade; if someone in a pai was caught committing the pertinent crime then
all ten households in that pai would be executed (Hu, 1987; Brook, 2005).15
An unwitting outcome of these draconian measures adopted by the Ming authorities had
likely predisposed the merchants toward raiding China’s coastal cities and towns that had
previously been actively engaged in maritime trade. As Wang Zhi—the head of a powerful
merchant-pirate once stated: “…if [the authorities] resume the customs in the ports of Zhejiang, and permit the people to trade with Japan, the pirates would not come again.…”
(Chao, 2005, p. 196). The rise of piracy along the China coast is demonstrated in Figure 2.
After nearly two centuries of less than 10 piratical attacks a year, from 1550 onwards the
number of piratical attacks rose sharply to over 90 per annum. In any case, the pirates now
shipped the commodities they plundered from the China coast to the many foreign markets
in Macau, Japan, Manila, Malacca, and even India for subsequent trading (So, 1975).
[Figure 2 about here]
But the imperial authorities’ attempt to suppress piracy “backfired”⎯to borrow Mann’s
(2011) word—when piracy unexpectedly surged. Afflicted by the piratical depredations and,
accordingly, the sharp deterioration in living standard in the distressed coastal communities,16 the imperial authorities softened its stance regarding the legitimacy of maritime trade
14
Due to the same reasons, the Qing dynasty (1644-1911) that replaced the Ming similarly imposed the
“sea ban” policy from approximately the 1650s to the 1840s.
15
This scheme was implemented by the Ming court in 1548, the primary goal of which was to curtail
smuggling in the coastal provinces in the southeast and banditry in the inland provinces. We know from
available evidence that this scheme was widely adopted in the coastal provinces of Fujian, Zhejiang, and
Guangdong after 1548, and was further extended to most other provinces by the end of the Ming dynasty
(Brook, 2005).
16
It is estimated that approximately 100,000 were killed and more than ten prosperous townships along
the coast were severely destroyed during the clash between the pirates and the imperial troops (Chao,
2005). Given that the majority of those who resided on the coast (especially those in the southeast) relied
heavily on maritime commerce for a livelihood, the suppression of maritime trade affected many severely
6
(von Glahn, 1996; Brook, 1998). In 1567, the year when Longqing became the new emperor,
he quickly legalized foreign trade by issuing licenses to Chinese junks to carry out trade.17
The result was a sharp plunge in piratical activities by the end of the 1560s, to a level similar
to that found in the two centuries before its initial surge. From then on until the end of the
Ming dynasty (1644), piracy hovered within a narrow range at a low level (Figure 2) in the
midst of a rapidly growing maritime trade (von Glahn, 1996; Frank, 1998).
3. Data
To test the hypothesis that the sharp rise in piracy in China in the 1550s was caused by
the interplay between the surge in demand for Chinese goods by the Europeans, on the one
hand, and prohibitions on the supply of these goods by the Ming authorities on the other,
we construct a unique panel data set that comprises all 33 coastal prefectures of Ming China
in cross-section and 270 years (1371-1640) in time-series.18 In order to avoid the inclusion of
social unrests that occurred in the last several years of the Yuan dynasty (1271-1368), we begin our analysis from 1371—two years after the founding of the Ming dynasty. By the same
token, we end our analysis in 1640—four years before the end of the Ming dynasty.
3.1. Piracy
Our dependent variable is measured by the number of piratical attacks made on a coastal
prefecture in each year (or decade).19 We define a piratical attack specifically as the act of
plundering the coastal towns and cities for goods that subsequently became diverted for
trade outside of China. The data are obtained from the Veritable Records of the Ming Dynasty
(Ming shi lu), which was officially compiled by the Ming court to record the edicts of the
emperors and memorials filed by the officials on events of national significance. In the Veritable Records, the time and place but not the scale or intensity of each piratical attack was
clearly recorded. In our sample, the mean of piracy is 0.10 per prefecture per year, with a
maximum of 23.
(So, 1975). For this reason, many local officials had repeatedly petitioned for the emperor to relax the “sea
ban” policy (Chao, 2005).
17
Trade with any country, except Japan, became legal. However, the licensing system broke down by the
1620s as the Ming dynasty went into decline. See von Glahn (1996, p. 118) for details on China’s licensed
foreign trade after 1567.
18
A prefecture was an administrative unit ranking below a province and above a county in Ming China’s
administrative structure which has remained valid to this day. We exclude the inland prefectures because
no piratical attacks occurred beyond the borders of the coastal prefectures.
19
The data are available on an annual basis. We sort them into decadal units for the sole purpose of checking the robustness of our estimations.
7
3.2. Trade Potential
Silk center. In light of the immense popularity of silk during the sixteenth century—a
product (both raw silk and silk products) over which China enjoyed a comparative advantage (Brook, 1998; Ma, 1998), our empirical analysis exploits the regional variation in the potential for silk trade (or raid in the case of piracy). Our use of the potential for silk trade is
premised on the historical fact that silk was the most sought-after good among the pirates:
“…the treasure that pirates sought was silk. When they found the workshops of silk production, they jumped for excitement…they even kidnapped Chinese women to secret spots and
forced them to weave silk” (Chao, 2005, p. 192).20 We thus employ a dummy variable indicating whether a prefecture was a major silk center with respect to the production of raw
silk and/or silk weaving to proxy for regional variations in trade (raid) potential. According
to Chinese historians, the imperial authorities had set up official bureaus in prefectures
where silk weaving was the mainstay of the local economy. Analogous in status to stateowned enterprises in contemporary times, these bureaus were made responsible for producing a variety of silk products for use by royal personnel and government officials. We count
only those prefectures where a silk-weaving bureau had been established as a silk center,
the geographic distribution of which is provided by Fan and Jin (1993). Of the 33 coastal prefectures in Ming China, 11 were silk centers. This variable is time-invariant because the geographic distribution of silk production rarely changed over time.
However, since China also exported many other commodities such as porcelain, tea, paper, and sugar (Brook, 1998), the silk center variable fails to fully capture the trade potential
of China’s coastal prefectures. To address this inadequacy we employ two additional measures to proxy for trade potential.
Historical port. Our first alternative measure exploits the variation in the endowment of
geography or specifically whether a prefecture had developed a port for conducting foreign
trade during the Song and Yuan dynasties (906-1368)—a time when China’s maritime commerce began to prosper. Given that China’s overseas trade that occurred prior to the Ming
dynasty was geographically concentrated in areas with port facilities like Quanzhou and
Ningbo (in the southeastern region), the potential for foreign trade was likely greater in prefectures with such facilities. In fact, the bulk of trading activities that occurred prior to the
advent of the great wave of piracy took place in close proximity to the ports in Quanzhou
(the island of Yuegang, for instance) and Ningbo (the island of Shuangyu) (Figure A1). Of
the 33 coastal prefectures in total in Ming China, 12 had a historical port.
20
The strong demand for Chinese silk is further evidenced by the fact that in the 1580s, after the “sea ban”
policy was removed, the import of Chinese silk contributed over 90% of customs revenue in Manila—the
major port for transpacific trade in Asia (Quan, 1986).
8
Urbanization. Another proxy for trade potential is the urbanization rate. As Acemoglu,
Johnson and Robinson (2002, 2005) have pointed out, in the absence of reliable GDP figures
urbanization rate is a reasonable proxy for the economic prosperity of pre-industrial societies. China is no exception: historians of China have indeed found that the urbanization
rates were closely correlated not only with the level of commercialization but also with exports (Skinner, 1977; Xu and Wu, 2000).21 We obtain the data on the urbanization rates or
specifically the share of the population living in the urban area in Ming China, computed at
the prefecture level, from Cao (2000).22 The mean urbanization rate among the coastal prefectures of Ming China is 11.3, with a maximum of 28.3.23
The spatial variations in trade potential and piratical attacks—the two key variables in
our analysis— are depicted in Figure 3A-D. Figure 3A shows the distribution of the number
of piratical attacks by coastal prefecture between 1550 and 1567, whereas Figures 3B-D specifically show the spatial variations in trade potential measured in terms of silk center, historical port, and urbanization rate. As one can readily see, piratical attacks were concentrated in regions with perhaps the greatest potential for maritime trade, such as the lower
Yangzi region and the coastal province of Fujian. In sharp contrast, few piratical attacks
were observed in north China, where the potential for maritime trade was limited.
[Figure 3 about here]
3.3. Controls
There are several variables that may be correlated with the incidence of piracy and thus
need to be controlled for. The first pertains to famine or negative economic shocks, which
are likely triggering forces of conflict (Miguel, Satyanath and Sergenti, 2004; Bai and Kung,
2011). Historians of China have indeed made the case that famine had led to piracy in Ming
China (So, 1975; Brook, 1998). To control for its possible effect on piracy we control for the
annual number of famines that had occurred in each prefecture.24
21
Bozhong Li (2000), an eminent Chinese historian, has found that regions highly developed in producing
silk, porcelain, and other exports in late imperial China were also more developed in their off-farm sectors
and accordingly had higher urbanization rates. Another eminent historian, Yiling Fu (1989), similarly
attributes the rapid development of the lower Yangzi region (particularly the prefectures of Suzhou,
Huzhou, Shengze, and Puyuan) during the Ming and Qing dynasties to the rapid development of a silk
industry and a flourishing commerce.
22
According to Cao (2000), urban areas (or cities) in the Ming dynasty included the national capital,
provincial capitals, prefectural seats, county seats, and commercial towns with a population above 1000.
23
This is slightly higher than the corresponding estimate for Western Europe (England, France, the
Netherlands, Portuguese, and Spain) of 10.1 in 1500 (Acemoglu, Johnson and Robinson, 2005).
24
To the extent that those suffering from famine tend to raid their neighboring prefectures rather than
their own, we employ the number of famines in adjacent, coastal prefectures as an alternative measure
and obtain similar results (not separately reported).
9
Population pressure represents another potential source of conflict (Malthus, 1798;
Bruckner, 2010). In the Chinese context, historians have documented a rapid increase in the
population during the latter half of the Ming dynasty (Ho, 1959; Cao, 2000). Forced upon by
pressure on the land, those on the coast had greater proclivities to turn to maritime trade
when permitted to do so, but were likely compelled to resort to piracy when trade was
banned. We thus need to control for the possible effect of China’s growing population pressure on the incidence of piracy. Given that China’s population data are available only at the
provincial level for our period of interest, we are only able to employ population density for
the periods of 1370s-1460s, 1470s-1550s, and 1560s-1640s, as proxies for population pressure.25
Last but not least, we control for the temporal variation in the regime’s military capability, proxied by the number of naval garrisons, as they were the most direct deterrent of pirate activities.26
The definition, sources, and descriptive statistics (mean and standard deviation) of the
main variables employed in this analysis are summarized in Table 1.
[Table 1 about here]
4. Empirical Strategy and Results
4.1. Flexible Estimates
We test our hypothesis by examining whether piratical attacks on prefectures with greater trade potential (relative to those with lesser trade potential) rose rapidly around 1550
when Ming’s imperial authorities stepped up its suppression of the growing trade activities,
and declined sharply after 1567 upon the removal of the “sea ban” policy. To provide a
benchmark, we estimate a fully flexible equation that assumes the following form:
pit = TPi × yeart + x'it + prefecturei + yeart + ε it
(1)
where pit stands for piratical attacks; TPi represents regional variations in trade potential
among the 33 coastal prefectures in Ming China, where trade potential is measured by
whether a prefecture was a silk center, had a historical port, and the prevailing urbanization
25
As with the case of famine, those in prefectures where population pressure was greatest were more
likely to raid on provinces other than their own. We thus also use mean population density in neighboring
coastal provinces as an alternative measure and obtain similar results.
26
Historians have indeed attributed the rise of piracy in mid-sixteenth-century China to the gradual
weakening of coastal defense (So, 1975; Chao, 2005).
10
rate; x'it is a vector of other covariates (famine, population density, and naval deterrence);
prefecturei denotes the prefecture fixed-effects capturing all time-invariant and prefecturespecific characteristics (such as geography, culture, and historical background); yeart denotes the time fixed-effects controlling for the common shocks to piracy in all the prefectures;
and ε it is the disturbance term. Our key interest lies in the set of interaction terms between
the variables of trade potential and a full set of year dummies, viz. TPi × yeart . These interaction terms are intended to capture the differential intensity of piratical attacks on prefectures
with varying degrees of trade potential on an annual basis. To the extent that the surge in
piracy was caused by the intensified suppression of trade, we expect the coefficients of TPi
to be significantly greater in magnitude between 1550 and 1567 but constant before 1550 and
after 1567.
The flexible estimation results are reported in Figure 4, in which we plot the coefficients
of the interaction term TPi × yeart within the 95 percent confidence intervals. Figure 4A reports the results obtained using the silk center as the proxy for TPi , and Figures 4B and 4C
report the results obtained using historical port and urbanization rate, respectively, as the
pertinent proxy. In all cases 1375 is treated as the reference year as no piracy was recorded
in that year. We obtain similar results using other years in early Ming as the reference year,
but we do not report them separately.
There are several important findings from Figure 4. Foremost is that, regardless of how
trade potential is measured, prefectures with greater trade potential do not necessarily experience distinctly more piratical attacks until approximately around 1550. After 1550, the
coefficient of the pertinent interaction term becomes decidedly larger, suggesting that the
intensified suppression of foreign trade by the imperial authorities from 1550 onwards had
possibly forced the merchants to commit piracy. This important finding also suggests that
the positive demand for Chinese goods by the Europeans after 1500 did not sufficiently give
rise to piratical raids, and piracy would not have reached the level observed had it not been
for the intensified effort of the imperial authorities to suppress the growing trade activities.
Second, the importance of trade suppression in partially triggering piracy is further borne
out by the fact that the differential intensity in piratical attacks between prefectures with varying trade potential that occurred around the 1550s disappeared by the end of the 1560s,
and remained at the low, pre-piracy-period levels all the way through the end of the Ming
dynasty.
[Figure 4 about here]
11
4.2. Main Results
To see if our flexible results are robust we estimate the same set of relationships with a
structured specification:
pit = TPi × Post1500 + TPi × Post1550 + TPi × Post1567 + x'it + prefecturei + yeart + ε it
(2)
where all variables are defined in the same way as in Equation (1). The only difference between Equation (1) and Equation (2) is that in (2) we interact TPi with only the three pertinent time dummies of Post1500 , Post1550 , and Post1567 . Our hypothesis would be substantiated if the coefficient of TPi does not change significantly after 1500 (the time when the
only shock was the sharp rise in trade demand from Europe), becomes positively significant
after 1550 (when the imperial authorities intensified the suppression of foreign trade), and
becomes negatively significant after 1567 (when the authorities abolished the “sea ban” policy).
[Table 2 about here]
The main results of Equation (2) are reported in Table 2. In Panel A, trade potential is
measured by silk center, and in Panels B and C by historical port and urbanization rate, respectively. In all cases we control for the prefecture fixed-effects and year fixed-effects, and
use robust standard errors clustered by prefecture to control for possible correlation within
a prefecture. Reported in column (1), our benchmark OLS estimate shows that there is no
significant difference in the number of piratical attacks between silk and non-silk centers
during the post-1500 period. This suggests that, left to itself Europe’s demand shock would
not have given rise to the surge in piracy in China.27 As expected, piratical attacks on the silk
centers did increase significantly after 1550 and decrease significantly after 1567. In terms of
magnitude, the size of the pertinent coefficient (of 1.501) indicates that the average number
of piratical attacks on the silk centers increased by 1.5 times more than the non-silk centers
after Ming China cracked down on foreign trade (i.e., Silk center × Post1550). Conversely, the
significantly negative coefficient of -1.513 for the Post1567 period suggests that after maritime trade was sanctioned piratical attacks on the silk centers had more or less returned to
the low level prior to its sharp increase (1.501-1.513 = -0.012).
In column (2) of Table 2 we include in the regression all other covariates, namely famine,
population density, and naval deterrence. Estimates on the three interaction terms are strikingly similar to those reported in column (1) in terms of both level of significance and
27
In light of the fact that the European traders first arrived in China in 1517, we also use 1517 as an
alternative cut-off year for the sharp rise in demand from Europe. The results are similar (and hence not
reported).
12
magnitude. To check the robustness of the results, we aggregate the yearly data into a decadal structure and perform the same estimations in column (3) (instead of 1567, 1570 is thus
used as the cut-off date of the removal of the “sea ban” policy). In addition, given that the
dependent variable is left censored with nearly 80% of the values being zero, we employ the
Tobit estimation (of the decadal data) and the results are reported in column (4). As we can
see the results of these additional checks are strikingly similar to what we have previously
found.
Panels B and C of Table 2 report the estimates using historical port and urbanization rate
as the alternative measures of trade potential, the results of which are also strikingly similar.
For example, the coefficients of historical port are very close to those of silk center in terms
of both magnitude and level of significance.28 Regarding the effect of urbanization, the estimates in column (2) suggest that after 1550, an additional percentage point of the urbanization rate increases the number of piratical attacks by a substantial 13 percentage points each
year, and decreases the number of piratical attacks at the same rate after 1567.29 These results
lend further support for the finding that the rising incidence of piracy in mid-sixteenthcentury China was caused to a minor extent by Europe’s growing demand for trade with
China but to a much larger extent by the crackdown of such activities by the imperial authorities.
4.3. Alternative Measure of Europe’s Trade Demand for Chinese Goods
In our main specification of Equation (2), the growth in demand from Europe for Chinese goods is proxied simply by the dummy variable Post1500. This measure is too crude,
however, since the number of Europeans arriving in China (stationing near the coast) kept
increasing after 1517, suggesting that their demand to trade with China was more likely a
gradual process. To measure Europe’s trade demand more accurately, we employ the total
tonnage of European ships that arrived in Asia after 1500 (“European ships”) as an alternative proxy. We choose this measure because no data exists on how many European ships
had actually arrived in China. Given this limitation, we deem European ships bound for
Asia a reasonable proxy for Europe’s trade demand with China primarily because maritime
trading routes connecting China to other parts of Asia were already well developed by the
early sixteenth century, around which time the Chinese merchants were found to have been
actively trading in many Asian ports such as Malacca and India (Findlay and O’Rourke,
2007). Decadal data on the total tonnage of European commercial ships that sailed to Asia
28
We also employ the number of years a historical port had acquired its status as an alternative measure
of trade potential and obtain similar results.
29
To fully measure the variations in trade potential, we also aggregate the three variables into a single
index by taking their first principal components. The results are similar to those using a single measure
and thus are not reported separately.
13
from 1500 onwards are obtained from de Vries (2003) (Figure 2).30 For estimation we specify
a regression equation of the following form:
pit = ship t × TPi + shipt × TPi × Post1550 + ship t × TPi × Post1570
+ x'it + prefecturei + yeart + ε it
(3)
where ship t denotes change in European demand to trade with China over time measured
by the total tonnage of European ships that sailed to Asia (in 1,000 tons). The interaction
term shipt × TPi identifies the effect due to the possible change in trade demand on the differential intensity of piratical attacks on prefectures of varying degrees of trade potential. The
triple interaction term shipt × TPi × Post1550 is intended to capture the effect of this change
on piracy after Ming China began to rigorously suppress foreign trade, whereas
shipt × TPi × Post1570 is intended to identify the change in this effect on piracy after the “sea
ban” policy was removed. We expect the coefficient of shipt × TPi to be significantly positive
between 1550 and 1570, but to turn significantly negative after 1570 according to our rehearsed logic. To comply with the decadal construction of the tonnage data, we similarly
aggregate the values of each variable in the regression on a decadal basis.
Table 3 reports the results of estimating Equation (3) based on the specification of Equation (2) and we found consistent results. For instance, the coefficient of shipt × TPi , which is
employed to identify the effect of the increase in trade demand on piratical attacks, is insignificant across all estimations. This suggests that, despite the sharp increase in European
ships that sailed to Asia from almost none before 1500 to an average of 40,930 tons during
the first half of the mid-sixteenth century, piratical attacks on the prefectures with greater
trade potential actually did not increase.
A more interesting finding is that shipt × TPi becomes significantly positive after 1550.
Using the result in column (2) of Panel A to illustrate, for example, an increase in the weight
of European ships that sailed to Asia by 1,000 tons increases the number of piratical attacks
on the silk centers by an average of 0.37 times per decade (or 0.037 times per year) more
than it does on the non-silk centers.31 While the marginal effect of trade expansion on piracy
may appear small, the total effect is huge (8.89 times per decade per prefecture or 0.889
times per year per prefecture),32 in the light of the fact that the total tonnage of European
ships that sailed to Asia increased by 24,036 tons between 1371-1550 and 1551-1570. Recall
earlier that the increase in the average number of piratical attacks on the silk centers during
30
We do not employ the number of European commercial ships to measure trade volume because vessel
size varied considerably during this period of trade expansion.
31
The average tonnage of a European ship that sailed to Asia between 1500 and 1640 was only 566 tons (de
Vries, 2003).
32
This is obtained by dividing 24,036 by 1,000 (tons), then multiplying it by 0.037.
14
1551-1567 was 1.5 times higher (per year per prefecture) than the corresponding increase in
the average number of piratical attacks on the non-silk centers, the increase in European
shipping tonnage in that period accounts for more than half of the difference in piracy between the two types of prefectures (0.889/1.5=0.589).
In contrast, the pertinent coefficient turns negative after 1570 (-0.36 in column (2) of Panel A), suggesting that the same increase in shipping tonnage will likely lead to a decrease in
piratical attacks of comparable magnitude on the silk centers. These results lend further credence to the hypothesis that the increase in the demand by the Europeans for trade with
China by itself has had no effect on piracy prior to the rigorous enforcement of trade restrictions by the Ming authorities around 1550.
[Table 3 about here]
5. Robustness Checks
5.1. Instrumented Evidence
Our key explanatory variable—a prefecture’s trade potential ( TPi )—is clearly endogenous. The main source of endogeneity is that our estimations may have omitted some variables that are correlated with both piracy and trade potential. A notable case in point is
that those living in regions with greater proclivities for engaging in maritime trade would
more likely possess the kind of “skill set”—such as knowledge of navigation—that would
facilitate piratical activities. There is also a concern about measurement error. Despite our
effort in constructing three measures to proxy for a prefecture’s trade potential, it would be
unrealistic to expect them to fully capture the variable in question.
To address these concerns, we exploit the exogenous variation in trade potential among
China’s coastal prefectures using the suitability of soil for growing mulberry leaves as the
instrumental variable for silk center. The rationale for employing this instrument is that
mulberry leaves is a critical input in the upstream production process of sericulture. Historical evidence suggests that, even up until the early 1920s, mulberry plantation was an
integral part of sericulture. Before the women could engage in silkworm-raising and weaving, the men had to plant the mulberry trees and stripped from the trees large quantities of
mulberry leaves during the intensive feeding periods when women were busy with five or
six daily feedings that went on around the clock (Bell, 1999). The complementary nature of
the production of mulberry leaves and cocoons, the highly perishable nature of mulberry
leaves and not the least the high cost of transporting them over long distances, together resulted in most silk production facilities locating in close proximity to sites where it was feasible to plant mulberry on a large scale (Fan and Jin, 1993). This depends crucially on the soil.
15
Mulberry crop is intolerant of waterlogging. The well-drained nature of loamy soil is
thus suitable for the large-scale plantation of mulberry (Dai, 1934). Given that soil texture is
exogenously determined and that, while its spatial distribution is likely correlated with
mulberry plantation and silk production, it should have no direct correlation with piracy,
and hence is a sound instrumental variable for silk center. We calculate the percentage of
loamy land in a prefecture based on the Soil Atlas of China, compiled in the 1980s by the Institute of Soil Science, Chinese Academy of Sciences. Because the texture of soil does not easily change over time even with human intervention (Rowell, 1994), contemporary data on
soil texture is a valid proxy for soil texture in Ming times.33
Table 4 tests the validity of our instrumental variable. In columns (1) and (2) we examine
the relationship between the percentage of loamy land and whether a prefecture was a silk
center. Both the OLS and Probit estimations for the 33 prefectures indicate a strong positive
relationship between the two variables. The exogeneity of loamy land is further examined in
columns (3) and (4). We first conduct a reduced-form analysis by regressing the number of
piratical attacks on loamy land in column (3), which shows that loamy land has a positive
and significant effect on the incidence of piracy. After including the silk center dummy in
the regression, however, the effect of loamy land disappears (column (4)). This suggests that
loamy land has no direct effect on piracy other than through the channel of silk centers.
[Table 4 about here]
We report the instrumented results in Table 5. Because the endogenous variable is the interaction term between silk center and the time dummies, we employ the interaction terms
of the share of loamy land and the same set of time dummies as instrumental variables. We
replicate what we did in Table 2 (the main results). The 2SLS (or IV-Tobit) results are consistent with those of the OLS (or Tobit) in terms of both direction and level of significance.
Moreover, the magnitude of the instrumented estimates of the two interaction terms are
greater than that of the OLS (or Tobit) results by about three times, suggesting that the endogenous silk center dummy has likely attenuated the estimated effect of trade potential on
piratical attacks.
[Table 5 about here]
33
In addition, mulberry also grows better when the soil pH-value lies somewhere between 6.5-7.5 (in a
range of approximately 4.5-8.5 in China). Our results (not reported) change little when we use the
percentage of loamy land and the percentage of land with a pH-value of 6.5-7.5 as instruments. However,
this result should be interpreted with caution since soil pH-value can be affected by human intervention.
16
5.2. Additional Controls
To the extent that trade was legalized in response to the heightened piratical activities as
some historians have suspected (von Glahn, 1996; Brook, 1998), removal of the “sea ban”
policy would be endogenous. To check if the dummy variable Post1567 is free from the
problem of reverse causality, we lag the number of piratical attacks. Doing so also helps alleviate the concern that the choice of location of a piratical attack may be endogenously determined more by “learning” from previous experiences (e.g., where it might be easier to
plunder) and less by trade potential.
These issues are addressed in Table 6, first by means of an OLS estimation (column (1)) ,
followed by the use of a generalized method of moments (GMM) approach based on Arellano and Bond (1991). To satisfy the conditions required by the “small t and large N” of GMM
estimation, we use decadal data (27 decades and 33 prefectures) to reduce the number of observations over time. The results remain robust.
Our difference-in-differences estimation requires that there were no other shocks occurring around the various cut-off points (1500, 1550 and 1567)—especially shocks that are correlated with piracy. To ensure that the periods selected for our difference-in-differences
analysis were not plagued by the “spillovers” of the prevailing political situation in Japan
during the mid-15th century (i.e., during the Sengoku or Warring States period), when the
lack of centralized political control then may have predisposed warlords and traders to turn
to preying along the China coast (Sansom, 1962), we add two interaction terms to identify
the effect (if any) of this possible Japanese influence.34
To capture the possible effect of the Japanese political turmoil, we interact trade potential
with the Post1467 time dummy. By the same token, to capture the effect of Japan’s reunification we interact trade potential with the Post1590 time dummy. Reported in column (3) of
Table 6 inclusion of Japan’s political influence does not change our main results.35
[Table 6 about here]
6. Conclusion
The Voyages of Discovery has been attracting increasing attention from economists interested in the long-term impact of this significant historic event on various dimensions.
34
Japanese piratical activities were effectively halted by Toyotomi Hideyoshi’s reunification of Japan in
1590, when he initiated the Sword Hunt and confiscated all weaponry from the peasantry. In particular, the
Daimyo were required to swear by the oaths to ensure that no seafarer would engage in piracy; those who
failed to comply would be deprived of their fiefdoms (Berry, 1989).
35
To save space we report only the results obtained using the yearly data.
17
While we are beginning to learn more about its economic impact on trade, population and
economic growth, standard of living and so forth, this understanding is confined, almost
exclusively, to the European context. Why did the same shock, which resulted in rapid trade
expansion and as a corollary economic growth in Europe, contribute, in sharp contrast, to
the largest maritime conflict in China? By analyzing a data set uniquely constructed from
historical archives, we show that the Chinese prefectures with greater trade potential were
indeed raided more frequently when maritime trade was suppressed most rigorously (between 1550 and 1567). Consistent with the logic of our hypothesized reasoning, these prefectures subsequently experienced a reversal of (mis)fortune consequent upon trade liberalization (after 1567). Together, these results suggest that the radically different outcomes between the East and the West was the combined result of the sharp rise in European demand
to trade with China and the latter’s reluctance to do the same.
18
References
Acemoglu, Daron, Simon Johnson, and James A. Robinson. 2002. “Reversal of Fortune: Geography and Institutions in the Making of the Modern World Income Distribution.”
Quarterly Journal of Economics, 107 (4): 1231-94.
Acemoglu, Daron, Simon Johnson, and James A. Robinson. 2005. “The Rise of Europe: Atlantic Trade, Institutional Change, and Economic Growth.” American Economic Review,
95 (2): 546-79.
Anderson, John. L. 1995. “Piracy and World History: An Economic Perspective on Maritime
Predation.” Journal of World History, 6 (2): 175-99.
Andrews, Kenneth R. 1964. Elizabethan Privateering: English Privateering during the Spanish
War, 1585-1603. Cambridge: Cambridge University Press.
Antony, Robert J. 2003. Like Froth Floating on the Sea: the World of Pirates and Seafarers in Late
Imperial South China. Berkeley, Calif.: Institute of East Asian Studies.
Arellano, Manuel, and Stephen R. Bond. 1991. “Some Tests of Specification for Panel Data:
Monte Carlo Evidence and an Application to Employment Equations.” Review of Economic Studies, 58(2): 277–97
Bai, Ying, and James K. Kung. 2011. “Climate Shocks and Sino-nomadic Conflict.” The Review of Economics and Statistics, 93(3): 970–81.
Bell, Lynda S. 1999. One Industry, Two China: Silk Filatures and Peasant-Family Production in
Wuxi County, 1865–1937. Stanford, CA: Stanford University Press.
Berry, Elizabeth. 1989. Hideyoshi. Cambridge: Harvard University Press.
Braudel, Fernand (translated by Siân Reynolds). 1992. The Mediterranean and the Mediterranean World in the Age of Philip II. New York: HarperCollins.
Brook, Timothy. 1998. The Confusions of Pleasure: Commerce and Culture in Ming China. Berkeley: University of California Press.
Brook, Timothy. 2005. The Chinese State in Ming Society. London: Routledge.
Bruckner, Markus. 2010. “Population Size and Civil Conflict Risk: Is there a Causal Link?”
The Economic Journal, 120 (544): 535–50.
Cao, Shuji. 2000. Zhong guo ren kou shi: Qing shi qi (Population History of China: Ming Dynasty).
Shanghai: Fudan da xue chu ban she.
Chao, Zhongchen. 2005. Ming dai hai jin yu hai wai mao yi (Sea Ban and Maritime Trade during
the Ming Dynasty of China). Beijing: Ren min chu ban she.
Chen, Maoheng. 1934. Ming dai wo kou kao lue (Investigation on Pirates in Ming Dynasty). Harvard-Yenching Institute.
Cordingly, David. 1996. Pirates: Terror on the High Seas, from the Caribbean to the South China
Sea. Atlanta: Turner Pub.
Dai, Licheng. 1934. Sang shu zai pei xue (Mulberry Plantation). The Commercial Press.
19
de Vries, Jan. 2003. “Connecting Europe and Asia: A Quantitative Analysis of the Caperoute Trade, 1497-1795.” In Global Connections and Monetary History, 1470-1800, edited
by Dennis O. Flynn, Arturo Giráldez, and Richard von Glahn. Aldershot, England; Burlington, VT: Ashgate.
Elvin, Mark. 1973. The Pattern of the Chinese Past. Stanford: Stanford University Press.
Fairbank, John K. 1968. The Chinese World Order: Traditional China’s Foreign Relations. Cambridge, Mass.: Harvard University Press.
Fan, Jinmin, and Wen Jin. 1993. Jiang nan si chou shi yan jiu (Research on the History of Silk in
the Lower Yangtze Delta). Beijing: Nong ye chu ban she.
Findlay, Ronald, and Kevin H. O’Rourke. 2007. Power and Plenty: Trade, War, and the World
Economy in the Second Millennium. Princeton, N.J.: Princeton University Press.
Frank, Andre Gunder. 1998. Reorient: Global Economy in the Asian Age. Berkeley, Calif.: University of California Press.
Fu, Yiling. 1989. Ming Qing shehui jingji bianqian lun (Development in Society and Economy during the Ming-Qing period). Beijing: renmin chuban she.
Gosse, Philip. 1932. The History of Pirates. New York: Tudor.
Hersh, Jonathan, and Hans-Joachim Voth. 2009. “Sweet Diversity: Colonial Goods and the
Rise of European Living Standards after 1492.” Working paper
Hillmann, Henning, and Christina Gathmann. 2011. “Overseas Trade and the Decline of
Privateering.” Journal of Economic History, 71 (3): 730-61.
Ho, Ping-ti. 1959. Studies on the Population of China, 1368-1953. Cambridge, Mass.: Harvard
University Press.
Hu, Zongxian. 1987 [1573-1620]. Chou hai tu bian (Atlas of Coastal Defense). Shanghai: Shanghai gu ji chu ban she.
Institute of Soil Science, Chinese Academy of Sciences. 1986. Zhongguo tu rang tu ji (The Soil
Atlas of China). Beijing: Di tu chu ban she.
Iyigun, Murat, and Watcharapong Ratisukpimol. 2011. “Learning Piracy on the High Seas.”
Working paper.
Landes, David S. 1998. The Wealth and Poverty of Nations: Why Some Are So Rich and Some So
Poor. New York: W.W. Norton.
Leeson, Peter T. 2007. “An-arrgh-chy: The Law and Economics of Pirate Organization.”
Journal of Political Economy, 115 (6): 1049-94.
Li, Bozhong. 2000. Jiang nan de zao qi gong ye hua, 1550-1850 (Early Industrialization in the
Yangzi Delta, 1550-1850). Beijing: Zhong guo she hui ke xue wen xian chu ban she.
Ljungstedt, Andrew. 1832. Contribution to an Historical Sketch of the Portuguese Settlements in
China. Macao.
Ma, Debin. 1998. The Great Silk Exchange: How the World Was Connected and Developed. In Pacific Centuries: Pacific and Pacific Rim Economic History since the 16th Century.
20
Edited by Dennis O. Flynn, Lionel Frost, and A. J. H. Latham. New York: Routledge.
38-69.
Malthus, T. Robert. 1798. An Essay on the Principle of Population as It Affects the Future Improvement of Society. London: J. Johnson.
Mann, Charles C. 2011. 1493: Uncovering the New World Columbus Created. New York: Alfred
A. Knopf.
Meng, Shaohua. 1999. Zhong guo zai huang shi ji (The History of Famines in China). Beijing:
Zhong guo she hui chu ban she.
Miguel, Edward, Shanker Satyanath, and Ernest Sergenti. 2004. “Economic Shocks and Civil
Conflict: An Instrumental Variables Approach.” Journal of Political Economy, 112 (August):
725-53.
Nunn, Nathan, and Nancy Qian. 2010. “The Columbian Exchange: A History of Disease,
Food, and Ideas.” Journal of Economic Perspectives, 24 (2): 163–88.
Nunn, Nathan, and Nancy Qian. 2011. “The Potato’s Contribution to Population and Urbanization: Evidence from a Historical Experiment.” Quarterly Journal of Economics, 126 (2):
593-650.
O’Rourke, Kevin H., and Jeffrey G. Williamson. 2002. “After Columbus: Explaining Europe’s
Overseas Trade Boom, 1500-1800.” Journal of Economic History, 62 (2): 417-56.
Pomeranz, Kenneth, and Steven Topik. 2006. The World that Trade Created: Society, Culture,
and the World Economy, 1400 to Present. New York: M.E. Sharpe.
Rowell, David L. 1994. Soil Science: Methods and Applications. Harlow, Essex: Longman Scientific & Technical.
Ptak, Roderich. 1982. “The Demography of Old Macau, 1555-1640.” Ming Studies, 15: 27-35.
Sansom, George. 1961. A History of Japan: 1334–1615. Stanford: Stanford University Press.
Skinner, G. William. 1977. The City in Late Imperial China. Stanford: Stanford University Press.
So, Kwan-wai. 1975. Japanese Piracy in Ming China during the 16th Century. Michigan State
University Press.
Starkey, David J. 2001. “Pirates and Markets.” In Bandits at Sea: A Pirates Reader. Edited by
C.R. Pennell. New York: New York University Press.
Sun, Guangqi. 1989. Zhong guo gu dai hang hai shi (The History of Voyage in Ancient China). Beijing: Hai yang chu ban she.
Travers, Timothy. 2007. Pirates: A History. Gloucestershire: Tempus Publishing.
von Glahn, Richard. 1996. Fountain of Fortune: Money and Monetary Policy in China, 1000-1700.
Berkeley: University of California Press.
Wang, Yuquan. 2007. Zhong guo jing ji tong shi: Ming dai jing ji juan (General History of Chinese
Economy: Ming Dynasty). Beijing: Zhong guo she hui ke xue chu ban she.
Wills, John E. 1979. “Maritime China from Wang Chih to Shih Lang: Themes in Peripheral
History.” In From Ming to Ch’ing: Conquest, Reign, and Continuity in Seventeenth-Century
21
China. Edited by Jonathan Spence and John Wills. New Haven, Conn.: Yale University
Press. 201-28.
Wills, John E. 1993. “Maritime Asia, 1500-1800: The Interactive Emergence of European Domination.” American Historical Review. 98 (1): 83-105.
Xu, Dixin, and Chengming Wu. 2000. Chinese Capitalism, 1522-1840. New York: St. Martin’s
Press.
22
250
200
150
100
50
0
1450 1460 1470 1480 1490 1500 1510 1520 1530 1540 1550 1560 1570 1580 1590 1600 1610 1620 1630
Europe-to-Asia ships
Europe-to-America ships
Figure 1. Growth of European Overseas Trade
Note: Volume of ships is measured in terms of tonnage (in units of 1,000).
Source: de Vries (2003).
100
90
80
70
60
50
40
30
20
10
Figure 2. Number of Piratical Attacks on Coastal China by Year
Source: Veritable Records of the Ming Dynasty (1368-1643).
23
1640
1630
1620
1610
1600
1590
1580
1570
1560
1550
1540
1530
1520
1510
1500
1490
1480
1470
1460
1450
1440
1430
1420
1410
1400
1390
1380
1370
0
A
B
C
D
Figure 3. Piracy and Trade Potential in Coastal Prefectures of Ming China
Notes: Piracy refers to the total number of piratical attacks between 1550 and 1567, enumerated based on
the Veritable Records of the Ming Dynasty. The location of a silk center is based on Fan and Jin (1993). The
location of a historical port is based on Sun (1989). Urbanization refers to the share of city population in
the early Ming period (1390s), and is based on Cao (2000). The map of Ming China (of the year 1391) is
based on “CHGIS, Version 4”, Harvard Yenching Institute (2007).
24
14
12
10
8
6
4
2
1636
1626
1616
1606
1596
1586
1576
1566
1556
1546
1536
1526
1516
1506
1496
1486
1476
1466
1456
1446
1436
1426
1416
1406
1396
1386
-2
1376
0
-4
Coefficient
Lower 95% CI
Upper 95% CI
A. Silk center × Year Dummies
14
12
10
8
6
4
2
1576
1586
1596
1606
1616
1626
1636
1586
1596
1606
1616
1626
1636
1566
1576
1556
1546
1536
1526
1516
1506
1496
1486
1476
1466
1456
1446
1436
1426
1416
1406
1396
1386
-2
1376
0
-4
Coefficient
Lower 95% CI
Upper 95% CI
B. Historical Port × Year Dummies
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
1566
1556
1546
1536
1526
1516
1506
1496
1486
1476
1466
1456
1446
1436
1426
1416
1406
1396
1386
-0.1
1376
0
-0.2
Coefficient
Lower 95%CI
Upper 95%CI
C. Urbanization × Year Dummies
Figure 4. Flexible Estimates of the Relationship between Trade Potential and Piracy
25
Table 1. Definition of Variables and Descriptive Statistics
Variable
Piracy
Silk center
Historical port
Urbanization
Famine
Naval deterrence
Population density
European ships
Loamy land
Definition
Number of piratical attacks on each coastal prefecture in each year.
Whether a prefecture was a major area for the
production of raw silk and/or silk weaving.
Whether a prefecture had developed a port for
foreign trade during the Song and Yuan dynasties (906-1368).
The share of population in each prefecture who
lived in urban areas in Ming period (1390s).
Number of famines in each prefecture per year.
Natural log of the number of naval garrisons in
each prefecture in each year.
Natural log of population density in each coastal province in the three periods of 1370s-1460s,
1470s-1550s, and 1560s-1640s.
Total tonnages of European commercial ships
that sailed to Asia (in 1,000 tons).
The percentage of loamy land in a prefecture.
Source
Veritable Records of
the Ming Dynasty
Fan and Jin (1993)
Sun (1989)
Cao (2000)
Meng (1999)
Atlas of Coastal
Defense
Cao (2000)
de Vries (2003)
The Soil Atlas of China
(1986)
26
Whole
sample
0.100
[0.802]
0.333
[0.471]
0.364
[0.481]
11.301
[5.902]
0.670
[1.017]
2.292
[0.986]
3.486
[0.948]
29.780
[36.178]
0.398
[0.413]
Mean [Standard Deviation]
1371150115511500
1550
1567
0.023
0.019
1.047
[0.177] [0.162] [2.831]
15681640
0.047
[0.288]
0.893
[1.214]
2.201
[1.020]
3.730
[1.123]
0.598
[0.835]
2.364
[0.947]
3.247
[0.699]
0.492
[0.697]
2.372
[0.950]
3.282
[0.688]
0.382
[0.687]
2.376
[0.950]
3.284
[0.687]
0
[0]
34.551
[3.298]
33.633
[0.964]
80.578
[28.110]
Table 2. Causes of Piracy: Main Results
The dependent variable is piracy
Panel A
Silk center × Post1500
Silk center × Post1550
Silk center × Post1567
R-squared
Panel B
Historical port × Post1500
Historical port × Post1550
Historical port × Post1567
R-squared
Panel C
Urbanization × Post1500
Urbanization × Post1550
Urbanization × Post1567
R-squared
Controls in each panel
Famine
Population density
Naval deterrence
Year fixed-effects
Prefecture fixed-effects
Observations
Yearly data
OLS
(1)
Yearly data
OLS
(2)
Decadal data
OLS
(3)
Decadal data
Tobit
(4)
-0.012
(0.020)
1.501***
(0.399)
-1.513***
(0.402)
0.24
-0.032
(0.024)
1.510***
(0.398)
-1.513***
(0.402)
0.24
-0.322
(0.245)
12.663***
(3.382)
-12.653***
(3.424)
0.49
-0.239
(2.583)
11.551***
(3.122)
-12.601***
(2.754)
0.39
0.0003
(0.019)
1.286***
(0.426)
-1.291***
(0.430)
0.23
-0.023
(0.024)
1.286***
(0.427)
-1.291***
(0.430)
0.23
-0.236
(0.243)
11.087***
(3.555)
-11.183***
(3.575)
0.45
0.147
(2.650)
10.843***
(3.164)
-12.673***
(2.788)
0.39
0.002
(0.002)
0.130***
(0.021)
-0.130***
(0.021)
0.25
0.0006
(0.002)
0.129***
(0.021)
-0.130***
(0.021)
0.25
0.006
(0.022)
1.089***
(0.177)
-1.097***
(0.177)
0.50
0.318
(0.248)
1.290***
(0.267)
-1.295***
(0.223)
0.42
Yes
Yes
Yes
Yes
Yes
891
Yes
Yes
Yes
Yes
Yes
891
No
No
No
Yes
Yes
8910
Yes
Yes
Yes
Yes
Yes
8910
Notes: Robust standard errors are clustered by prefecture and reported in parentheses. In
columns (3) and (4) 1570 (instead of 1567) is employed as the year in which the “sea ban”
policy was removed. *** Significant at the 1% level.
27
Table 3. Causes of Piracy: Alternative Measure of Europe’s Trade Demand
The dependent variable is piracy
OLS
(1)
OLS
(2)
Tobit
(3)
-0.002
(0.006)
0.367***
(0.098)
-0.364***
(0.100)
0.48
-0.007
(0.007)
0.366***
(0.099)
-0.360***
(0.099)
0.48
0.032
(0.073)
0.327***
(0.092)
-0.345***
(0.075)
0.39
0.001
(0.005)
0.321***
(0.103)
-0.322***
(0.104)
0.46
-0.006
(0.007)
0.321***
(0.103)
-0.318***
(0.104)
0.46
0.034
(0.075)
0.314***
(0.093)
-0.341***
(0.075)
0.39
R-squared
0.001
(0.001)
0.032***
(0.005)
-0.032***
(0.005)
0.50
0.000
(0.001)
0.032***
(0.005)
-0.032***
(0.005)
0.49
0.013*
(0.007)
0.038***
(0.008)
-0.044***
(0.006)
0.24
Controls in each panel
Famine
Population density
Naval deterrence
Year fixed-effects
Prefecture fixed-effects
Observations
No
No
No
Yes
Yes
891
Yes
Yes
Yes
Yes
Yes
891
Yes
Yes
Yes
Yes
Yes
891
Panel A
European ships × Silk center
European ships × Silk center × Post1550
European ships × Silk center × Post1570
R-squared
Panel B
European ships × Historical port
European ships × Historical port × Post1550
European ships × Historical port × Post1570
R-squared
Panel C
European ships × Urbanization
European ships × Urbanization × Post1550
European ships × Urbanization × Post1570
Notes: Robust standard errors are clustered by prefecture and reported in parentheses. In all columns 1570 (instead of 1567) is employed as the year in which the “sea ban” policy was removed.
* Significant at the 10% level; *** significant at the 1% level.
28
Table 4. Testing the Validity of the Instrumental Variable
The dependent variable is
Loamy land
Silk center
Silk center
Piracy
Piracy
OLS
Probit
OLS
OLS
(1)
(2)
(3)
(4)
0.178*
(0.084)
1.183*
(0.696)
17.032**
(7.606)
Yes
Yes
Yes
0.49
33
Yes
Yes
Yes
0.47
33
Yes
Yes
Yes
0.47
33
Silk center
Famine
Population density
Naval deterrence
R-squared
Observations
13.414
(8.168)
20.340**
(9.390)
Yes
Yes
Yes
0.55
33
Notes: In columns (3) and (4) the dependent variable is the total number of piratical
attacks on each prefecture for the period 1371-1640. Figures for famine, population
density, and naval deterrence are the average values for 1371-1640. Robust standard
errors are reported in parentheses. * Significant at the 10% level, ** significant at the
5% level.
29
Table 5. Causes of Piracy: Instrumented Results
The dependent variable is piracy
Silk center × Post1500
Silk center × Post1550
Silk center × Post1567
Famine
Population density
Naval deterrence
Year fixed-effects
Prefecture fixed-effects
Observations
Yearly data
Yearly data
Decadal data
Decadal data
2SLS
2SLS
2SLS
IV-Tobit
(1)
(2)
(3)
(4)
0.758
(5.029)
37.884***
(9.087)
-38.179***
(8.800)
Yes
Yes
Yes
Yes
Yes
891
1.083
(5.684)
32.368***
(6.827)
-38.327***
(6.071)
Yes
Yes
Yes
Yes
Yes
891
0.082
(0.073)
4.456***
(1.618)
-4.481***
(1.618)
No
No
No
Yes
Yes
8910
0.077
(0.080)
4.497***
(1.672)
-4.535***
(1.646)
Yes
Yes
Yes
Yes
Yes
8910
Notes: Robust standard errors are clustered by prefecture and reported in parentheses. In all
columns silk center is instrumented by the percentage of loamy land. In columns (3) and (4)
1570 (instead of 1567) is employed as the year in which the “sea ban” policy was removed.
*** Significant at the 1% level.
30
Table 6. Causes of Piracy: Additional Controls
The dependent variable is piracy
Decadal data
OLS
(1)
Decadal data
GMM
(2)
Yearly data
OLS
(3)
-0.348
(0.272)
14.760***
(5.275)
-14.688***
(5.253)
-18.391
(21.302)
32.808***
(5.321)
-21.745***
(6.033)
0.020
(0.014)
1.509***
(0.398)
-1.568***
(0.407)
-0.213
(0.264)
12.391**
(5.057)
-12.455**
(5.032)
94.466
(68.195)
11.083**
(5.625)
-14.526**
(6.535)
0.017
(0.014)
1.286***
(0.426)
-1.316***
(0.445)
-0.156
(1.216)
2.037**
(0.980)
-3.051
(3.387)
0.002*
(0.001)
0.129***
(0.021)
-0.137***
(0.022)
Yes
Yes
Yes
No
No
Yes
Yes
Yes
858
Yes
Yes
Yes
Yes
Yes
No
Yes
Yes
8910
Panel A
Silk center × Post1500
Silk center × Post1550
Silk center × Post1567
Panel B
Historical port × Post1500
Historical port × Post1550
Historical port × Post1567
Panel C
Urbanization × Post1500
Urbanization × Post1550
Urbanization × Post1567
Controls in each panel
Famine
Population density
Naval deterrence
TPi × Japan Post1467
TPi × Japan Post1590
Piratical attacks t-1
Time fixed-effects
Prefecture fixed-effects
Observations
0.004
(0.022)
1.237***
(0.260)
-1.229***
(0.254)
Yes
Yes
Yes
No
No
Yes
Yes
Yes
858
Notes: Robust standard errors are clustered by prefecture and reported in parentheses.
* Significant at the 10%; ** significant at the 5% level; *** significant at the 1% level.
31
Appendix
Figure A1. Asia and Ming China in the Sixteenth Century
32