Trend in the Korean Stature, 1547-1882[*]
Myung Soo Cha and Young-Jun Cho
We estimate stature trend in Korea from 1547 to 1882using eleven sets of height records createdby judicial and military authorities. Stature figuresare given in terms of different yardsticks of unknown lengths and sometimes in round figures, making it difficult to estimate height precisely. Direction of height shift over several decades spanning different parts of the three and half centuries can be determined, however, which was either stagnant or downwards, but never upwards. Slaves had height advantage over free peasants, male adults were taller in northern than in southern provinces, andthe effect of smallpox on height was ambiguous.
Keywords: stature; pre-colonial Korea; conscription; postmortem
JEL classification: N35, N15
In the past half century, adult male height grew by roughlyten centimeters in South Korea, while North Korea hardly achieved any gain in stature. In both northern and southern provinces, Korean males became taller by more than two centimeters under Japanese rule from 1910-45.[1] The distinctrecords of height growthare consistent with widely differenteconomic performance achieved by the three regimes:colonial and South Korea saw per capita output grow 2.3% per year from 1911-40 and6.2% from 1960/2-2008/10, respectively, while living standardsat best stagnated in North Korea.[2] Multiplying the growth rates with Steckel’s estimate of one percent growth in per capita output increasing male adult height by 0.0397 centimeter yields predicted height growth, which are reasonably close to observed increase in stature: 2.6 (=0.0397×2.3×29) centimeters for colonial Korea, 11.8 (=0.0397×6.2×48) centimeters for South Korea, and nil for North Korea.[3]
Did Koreans grow or shrink in the centuries preceding the arrival of Japanese rulers? Did the twentieth century stature growthrepresent extension of pre-colonial trend or a sharp break with the dynastic past? Given the close association between height and income growth in twentieth century Korea, the answers to these questions is likely to provide important clue about the growth performance of pre-colonial Korea, which is difficult to measure given the paucity of reliable data on population and output. Whether Koreans became better or worse off in the centuries preceding the imposition of Japanese rule bears on the origins of Korea’s modern economic growth, which remains an issue charged with emotion and debated passionately in the Korean historiography.[4] This article is an attempt to throw light onwhat triggered Korea’s modern economic growth by estimating height trend in Korea from the mid-sixteenth to late nineteenth century.
To estimate height trend in mid- and late dynasticKorea, we useelevensets of archival resources created by judicial and military authorities and providing stature information for more than four thousand adultsborn from 1547-1882. We begin by familiarizing readers with the data sources in the first section. Eleven height datasets were compiled from the sources, whichare different from each otherin terms of not only yardstick used to measure stature, but also the nature of non-stature information provided. Hence, the second section analyzes each of thedatasets separately using dummy variable regression to estimate fragmentedheight trends, each of which isseveral decades long and scattered all over the three and half centuries preceding the beginning of Japanese rule. In the third section, we integrate the findings made in the second section to conclude that stature is likely to have trended downwards, rather than upwards in Chosŏn Korea. The final section concludes by discussing the implications of the height contraction in pre-colonial Korea on what triggered modern economic growth in Korea.
Conscription andAutopsyin Dynastic Korea
Wars with Japan (1592-8) and China (1636-7)caused serious damage to people and land in Korea, fundamentally changingthe system of government in Korea,fiscal structure in particular. Before the wars, the dynastic government collected taxes in kind and imposed compulsory military service on able-bodied males of free status. The wars having decimated labor force and made it virtually impossible to operate the tribute system, the bureaucracy needed to draw on markets to achieve fiscal extraction and military mobilization. The postwar government collected taxes in rice, a money commodity, which could be used by the government to obtain items it needed on the market. Rather than conscripting peasants by compulsion, the government hired career soldiers, who were remunerated with cotton cloth, the other type of money commodity, collected from peasants as military tax. A new military organization, hullyŏdogam (訓鍊都監), was created recruiting career soldiersnot only from free peasants, but also from slaves, who had—together with the aristocracy, known as yangban(兩班)—hitherto been exempt from military service. This scheme did not completely replace thetraditional system, which was re-organized into sogogun(束伍軍), comprising peasants required to pay cloth tax in ordinary times, but supposed to serve as soldiers in the case of hostilities. The traditional system of national security was finallyrevamped in the wake of modernization drive launched in 1894 to be replaced with a standing army staffed by career soldiers.
[Table 1 about here]
The military authorities of dynastic Koreameasured the bodies of soldiers, presumably with a view toleaving adescription of their physical characteristicsto be used to track down escapeesin the event of desertion, rather than checking whether they were fit for military service: as discussed in the following section, there is little indication of minimum height requirement applied in recruiting soldiers. We were able to obtain eight sets of physical examination records created by military authorities in different years from 1595 and 1901, which are presented in Table 1. The first six rows from the top of Table 1 list records generated before 1894 in an effort to organize sogogun for different regions; physical examination records for the standing army defending the capital, hullyŏndogam, is unfortunately unavailable. On the two rows at the bottom of Table 1 are given the records of two different components of the reformed army, regular forces created after 1894 to defend the royal palace and northwestern part of Korea.
Stature figures found in the conscription documents are, without exception, in the traditional measure of length,known as ch’ŏk (尺), in Korea.[5] As several studiesindicated, one ch’ŏkrepresented variable lengths, depending on period, region, and purpose of measurement.
[Table 2 about here]
At the end of the nineteenth century, Japanese observers found five different types ofch’ŏkin frequentusein Korea for distinct purposes, the shortest and the longest of which being equivalent to roughly twenty to forty eight centimeters, respectively (Table 2). In addition to the fivecommontypes shown in Table 2, a number of different kinds ofch’ŏkapparently existed for localized use. The eight sets of military records listed in Table 1 were created applying different yardsticks, some of which possiblycorresponding with none of the five measures listed in Table 2 (see Figure 8 below).
Table1suggests that conscription occurred more frequently in the sixteenth and seventeenth centuries, when Korea fought wars with Japan and China, than in the following centuries. Evidence also suggests that significant changes occurred after 1700 in the way how conscription was implemented. Not only was height information given in round numbers, but also wecame across several sets of military stature record created, where an identical number was givenasthe height of a number of different soldiers. Some of thesewere created even usingforms, where fixed height figures had been printed. These may be interpreted as indication of national defense having become a less pressing issue, leading the government to spend less time and resources for security in the eighteenth and nineteenth centuries, when peaceprevailed. They may also reflect they system of government breaking down orthe military authorities seeking for individuals of certain stature. In any case, it is clear that such materialscannot serveas a source of height information.
Military rosters providing usable height figures are unavailable from the late eighteenth to late nineteenth centuries, which prompted us tosearch for non-military sources of height information. One such archival materialwe come across waspublic documents reporting the details of shipping tax rice collected by provincial governments to the capital, Seoul. The description included names, ages, and physical features, including height, ofsixty twoworkers employed in ships carrying ricedeparting from the southern coast.
The other non-military source of height informationhad to do with crimes. Judicial authorities carried out post mortem investigationon convictsdying in custody and personskilled byaccidents, which were mostly related to assault. Autopsy reportswere created following the examinations, detailingthe stature of corpses, causes of death, and places of origin. Some reports specify thetype of ch’ŏkapplied as well, revealing that the judicial authorities differed little from the military of dynastic Korea in being unable to impose a unified procedure of body measurement on officials stationed in different regions.
[Table 3 about here]
Table 3 lists three differentarchival sources of height information as generated by postmortem examination. The first set includesthe summary of postmortem examinationof bodies of accidental death, which are found in reportssubmittedby provincial governors to the central governmentin the eighteenth century. The second and third sources, providing height data for individuals born in the nineteenth century, give full results of autopsy of murdered individuals and convicts dying in jail, respectively. While the height observations available from judicial documents far outnumber that found in public employment documents, the amount of height information provided by the two non-military sources are dwarfed by that found in military rosters.
Postmortem reports have several advantages over military records of physical examinations as a source of height information. One is that height observation is available not only for males, but also for females. Second, while military rosters give height observations made in a particular year for persons born in different years preceding the year of physical examination, autopsy is carried outin at diverse points in time on bodies of persons born in different years, which allows estimation of impact of age on stature along with stature trend. Finally, height information given in individualized autopsy reports is likely to be more accurate than those found in military documents, which were created by measuring a large number of males within a limited time span. Officials usually measured the length of a corpse twice and in exceptional cases upto five times to insure precision, recording the result down to two decimal places. This compares with height figures found in military documents, which were frequently rounded off to the nearest whole number (in the eighteenth century) or tenth (in the sixteenth and seventeenth centuries).
It should however be noted that theseadvantagesmay be offset by the possibility that shocks causing death may deform bodies, and that corpsesundergo transformation. Although in many cases examiners could obtain age information on murdered persons from their relatives, sometimes ages of victims had to be inferred from appearance, hence could be imprecise.
Analysis of Individual Datasets
From the archival sources introduced in the preceding section, we put up eleven datasets, each of whichrecording height information in terms of distinct unit of measurement. Had they differed only in terms of unit of measurement, conversion ratios between different yardsticks may be estimated simultaneously with stature trend. They provide different amount and type of non-stature informationas well, which makes it more efficient to adopt a two-stage approach, first analyzing individual dataset separately and then estimating long run trend and conversion ratios using the results from the first stage.
Soldiers and persons dying of accidents in Chosŏn dynasty included young persons, who had yet to achieve their full height. As Sinclair and Dangerfield(1998: 33) state, “Although noticeable growth in height stops at about 18 years in the female and 21 in the male, the vertebral column can continue to grow up to about 30 years of age.” Thus, one might ideally like to estimate stature trend from height of persons older than twenty nine, a demarcation, which carries the cost of reducing the number of usable observations substantially. To maximize sample size, wedefine adults as personsolder than twenty four, because, as Sinclairand Dangerfield(1998: 33) continues, the increase in height due to the growth of vertebral column is “little more than 3-4 mm, and the difficulties of assessing measurements of so small an amount will at once be apparent.” In addition, our definition of being adultwas foundas consistent with the result of the statistical analysis of height data taken from autopsy reports, which indicated that the effect of being in the age group from 20-24 on stature was insignificant.[6] It is also worth noting that an analysis of Chinese height data taken from postmortem records found that Chinese males stopped growing after twenty two.[7]
- The Sixteenth Century: Roster of P’yŏng’ando Garrison Command
Listing soldiers in reserve forces organized to defend P’yŏng’an province located in the northwestern corner of the Korean peninsula from Hideyoshi’s invading forces (see map of Korea provided as appendix), this source includes height information for male adults of both peasants and slave status. In addition to height and birth year, the place of origin, presence of pockmark, social status, muscle strength, and position are specified. The soldiers were from four different counties of P’yŏng’an province, and inspectors identified three different types of face,i.e. pockmarked, lightly pockmarked, and smooth. Social status refers to the distinction between free peasants and slaves, excluding the privileged class, i.e.yangban. Muscle strength is indicated by the maximum weight each soldier was able to lift. Some of the soldierswere assigned to specific positions, which included archer, musketeer, andspear bearer, among others.
[Figure 1 about here]
Figure 1 presents the distribution of height of three hundred and thirtythree soldiers of known status and aged older than twenty four, excluding nine height observations for males born before 1547, which appeared as outliers. Height observations, mostly given to one decimal place, appear as approximately normally distributed, although one observes substantially higher frequency of observations at the value of 7.5 and 8.0. While the concentration of observation at 8.0 probably resulted from assigning the value eight to all soldiers with height equal to or taller than 8, the frequent occurrence of 7.5 may reflect height heaping. Finally, the distribution suggests minimum height requirement was not applied in organizing the reserve unit.
Log-transformed height of males older than twenty fourwas regressed on four dummy variables indicating 1) birth decade, 2) place of origin, 3) pockmark, and, finally, 4)social status. Log-transformation of the dependent variable allows on to measure the impact of the explanatory variables on height in term of per cent, which appears as a more useful procedure thanevaluatingthe effect in terms of ch’ŏk, the length of whichremains unknown. Dummy variables indicating either muscle strength or position were not used, because they are very likely to be correlated with disturbance term, hence cause endogeneity bias.
[Table 4 about here]
Given that height values higher than eight are non-existent, we tried both ordinary least squares and truncated regression, whichyielded very similar results as shown by the two panels of Table 4.[8] Soldiers with smooth and pockmarked faces did not differ significantly in terms of height. Second, neither was status—free orslave— found to affect height. Finally, the estimated coefficients of associated with dummy variables representing birth decades indicatedstature trended downwards in mid-sixteenth century P’yŏng’an. Although anthropometric studies tend touse birth decade dummies to estimate stature trend, we also tried birth year dummies. The height calculated with the coefficient estimates of birth year dummies turned out to include a linear trend, which was statistically significant and negative, confirming the conclusion based on birth decade dummies. It is worth noting that estimation with birth year dummies indicated slaves were significantly taller than ordinary peasants.
- The Seventeenth Century: Regimental Troops in Ch’ungch’ŏngand Cheju Provinces
Three sets of military rosters are available for the seventeenth century,including the listof males organized intotwo distinct sogogununits in Ch’ungch’ŏng, a province located in the middle part of the Korean peninsula, and onereserve unit in Cheju, an island off the southern coast of the Korean peninsula.Two separatedatasets werecreated for Ch’ungch’ŏng(Ch’ungch’ŏng A and B hereafter), which include informationon adult males born from 1619-54 andthose born from 1629-72, respectively. Social status of soldiers isspecified for about four fifthsof adult males in Ch’ungch’ŏng A, but only for one fifth of those in Ch’ungch’ŏng B. With the exception of a small number of soldiers,Ch’ungch’ŏngA and B group faces into fivetypes, which includedbeing smooth and four different degrees of being pockmarked. The dataset built from the Cheju roster includes height information for male adults born from 1625-59, recording social status and defining faces into four different types (smoothand pockmarked in three different degrees)formost the soldiers.