Home Contact us Site Map

Cette page existe uniquement en anglais

Home > Activities > Committees >Historical Demography (1997-2001) > Seminar Report

Seminar on Population and the Economy: from Hunger to Modern Economic Growth

Toyonaka, Japan, 7-10 January 1997
Co-organized by the IUSSP Committee on Historical Demography and
the International Association of Economic History


The seminar was co-sponsored by IUSSP's Historical Demography Committee and the International Association of Economic History. Its objective was to provide a forum for the presentation and discussion of findings concerning the relationship between population and economic change in pre-industrial Europe and in non-European regions, both in the past and more contemporary times. The various papers presented form part of a report, and a published volume, presented and discussed at the Twelfth International Economic History Congress which took place in Seville, Spain in August 24-28, 1998.

Its organizers hoped to generate research that would bring us closer to filling existing gaps in our knowledge concerning Malthusian demographic-economic dynamics. While considerable light has already been shed on these dynamics, there still remain large gaps in our understanding of the mechanisms that underlie them. At the macro-level, questions remain concerning the mechanisms by which economic forces conditioned demographic behaviour and in turn how population constrained economic growth, in both the pre-industrial European and non-European contexts. At the micro-level, we know little of how individuals and households reacted to various forms of external stress, both in the short and long-term. Although the role of migration as both a short and long-term adjustment mechanism to population pressures and economic stress has recently been emphasized in the literature, we still have limited knowledge concerning this mechanism. Similarly, it is commonly held that the relationship between population responses and economic changes attenuated as a result of demographic and economic modernization, yet the timing, direction and pace of this process are still not fully understood.

The papers presented in this seminar deal with many of these issues. They make use of long-term economic and demographic data series, original methodologies that include both macro and micro-level analytical approaches, and provide cross-cultural and comparative perspectives. The seminar consisted of seven sessions, covering such issues as household context and mortality outcome, migration as demographic mechanism, long and short-term demographic-economic dynamics and the mediating role of demographic regimes. Household and mortality

In the first paper presented, 'Infant Mortality, Child Neglect and Child Abandonment in European History: A Comparative Analysis', Katherine Lynch explores whether conscious mortality 'strategies ' operated at the household level in the European past. By 'strategies' the author refers to parents' conscious regulation of their fertility through infanticide, child abandonment and infant and child neglect. She argues that when the social-cultural context of the period under study is taken into account it is difficult to conclude that these practices were 'normal or accepted ' by the general population or in any sense perceived as 'strategies'. She brings to bear numerous contextual and evidentiary factors. Among them, the importance and pervasiveness of Christian values which renounced infanticide and child neglect and abandonment. Others include the long history of foundling homes, established for the purpose of assisting women in dire circumstances deal with an unwanted child, and women's perception of them as temporary surrogate parents during 'critical life' situations. Likewise evidence from court documents is used to reveal that throughout the centuries the culprits of infanticide were mostly young, unmarried, poor women. That the women were treated leniently by juries reveals just how 'unnatural' these acts were believed to be. Similarly, she finds child neglect in the household did not follow any pattern or cultural norm that led parents to 'systematically target specific types of children for neglect'. Finally, she points out that for such behaviours to be deemed 'strategies' one must assume individuals believed that their environment remained stable over time. In a highly variable epidemiological regime, such an assumption is untenable.

In their study, 'Effects of Household Structure and Relationship on Mortality in Early Modern Japan: Evidence from Two North-eastern Villages', Noriko Tsuya and Satomi Kurosu examined the patterns and covariates of mortality within the household and in the context of kin relationships in two agricultural villages in North-eastern Japan in the 18th and 19th centuries. They found that both family structure and community wide factors affected individuals' mortality outcomes differently according to their age, sex, status in the household, kin composition within the household and household socio-economic status. For example, girls under the age of 15 experienced higher probabilities of dying than boys if they came from poor households. Boys, on the other hand, experienced greater survivorship probabilities with increasing number of servants, signalling how boys benefited more than girls from increased household wealth. Adult males aged 15 to 54 benefited much more from the effects of owning land than did their spouses, attesting to differential status and obligations within the household by gender. Finally, they report that as both men and women aged their survivorship increasingly depended on family/household structure, particularly kinship ties and marital status characteristics.

In their paper, 'Price Fluctuations, Family Structure and Mortality in Two Rural Chinese Populations: A Comparison of Peasants and Serfs in 18th and 19th Century Liaoning', Cameron D. Campbell and James Z. Lee investigate how household organization mediated the demographic effects of one type of economic stress (grain price fluctuations) and those of community setting. They were particularly interested in testing the hypothesis that the most privileged household members, those who received a disproportionate share of household resources in good times, would be the most affected in bad times. A second goal was to explore how household organization mediated the demographic effects of community setting. The authors found that despite differences in overall mortality outcomes between the two villages, which they attribute to differences in the disease environment, mortality differentials by household context were similar. Within household mortality outcomes varied according to age, sex, marital and kinship status. For example, female but not male orphans experienced higher mortality, as did widowed young women, married women with no male children and older widowed men. As they hypothesized, mortality differentials within households narrowed as prices increased as did inter-household mortality differentials, suggesting an effect of household socio-economic status. Finally, they found that household mortality differentials present in both villages were considerably stronger in the most remote village, highlighting the significance of household context as a source of individual variation in access to resources.

In her presentation, 'Remarks On What Can and Can Not be Known About Mortality in Early Modern Japan', Laurel Cornell addressed the following questions: why investigators should be interested, what period is of relevance, what can be known, and what can be done concerning the study of mortality in early modern Japan. She posited the uniqueness of Japan's mortality transition in terms of timing and relationship to economic changes as some of the factors making its study relevant. Second, she described Japan's economic and social history and indicated the period of relevance for the proper study of Japan's mortality transition. Third, she provided information on the types of data sources available and what can be gleaned from them. For instance, a characteristic of Japan's annual census registers (the source that is available to investigators) is that it does not contain information on infant deaths. Thus, the calculation of certain demographic indicators is not feasible and certain estimation techniques that applied elsewhere should be discarded. Finally, she argued that although the type of data available prevents the estimation of certain indicators, investigators can fruitfully carry out analyses of overall mortality trends.

In the discussions that followed a few important issues stood out. Among them the issue of cross-cultural and cross-class comparisons, data reliability and methodology. In the case of the Lynch study, it was noted that her analysis would have benefited by a more detailed comparison of the patterns of infanticide and child abandonment between Europe and Asia as well as by a more in-depth examination of differentials by social class. The Tsuya/Kurosu and Lee/Campbell papers generated lively discussions concerning data reliability and methodologies. The demographic data series employed in both cases lacked information on infant and child mortality, thus some of the conclusions drawn were questioned. For instance, gender differentials in expectation of life differed considerably between the Chinese and Japanese cases as did overall life expectancy when compared to the European case. The point made by Cornell concerning the reliability of certain estimates was highlighted in these two cases. Finally, it was noted that researchers must be very careful about employing highly sophisticated statistical tools to data that may be suspect. Migration in demographic contexts

In this first presentation, Massimo Livi Bacci posited the hypothesis that 19th century European emigration served as an escape route from Malthusian population pressures. As background, he cited the fact that approximately 50 million Europeans left the continent between the mid-19th century and World War II. Between 1860 and 1930, in Italy alone, the number of departures amounted to 25 million, with a net loss of 8 million individuals. He noted that if this emigration had not taken place Italy's population today would be larger by 25 million. Equally significant was that this escape route was chosen by some countries but not others. For example, while Italy exhibited a propensity for large scale emigration, France did not. He proposed then that the task for demographers is to unearth the factors behind this phenomenon and the reasons that explain why they played themselves out in some settings but not others. Possible starting points include investigations concerning the ability of native manufacturing sectors to absorb surplus rural labour, factors leading to increases in rural populations, and the role played by American demand for labour.

In his paper 'Migration, Marriage and Population in the Netherlands, 1500-1900' Jan de Vries considers the role of migration in the patterns of population growth, urbanization and nuptiality in the Netherlands during the 1500-1900 period. He divides the RePUBLIC's population history into three periods characterized by rapid population growth and urbanization (1500 to 1650-75), population stagnation and de-urbanization (1650-75 to 1815-50) and resumption of population growth and re-urbanization (1815-50 to 1900-14). According to the author, migration played a significant role in these population trends. He describes a process whereby migrant sex, age and marital status composition and type of migration (permanent or temporary) exuded considerable influence. During the first period a large proportion of migrants were family units and single males, while during the second, flows were dominated by single females. A considerable component of the migratory streams were composed of sojourners. Effects of changing migrant composition and types of migration were manifested in female age at first marriage and marriage rates. These changing nuptiality patterns contributed in turn to fertility decline. The role of migration in the latter period's renewed population growth lay in the breakdown of labour migration. The author claims that, ironically, it was a reduction in the Dutch economy to a size compatible to its domestic labour force that allowed the population to achieve positive rates of natural increase.

In the following paper, 'Population Growth and Population Regulation in Nineteenth Century Rural Scotland', Michael Anderson addresses the question of whether the high emigration rates and depressed nuptiality of the second half of 19th century rural Scotland were a response to over population, the result of a breakdown of population control mechanisms during the latter part of the 18th century. He describes an overall pattern in which the rural population experienced rapid growth prior to 1821 and later a steady decline throughout most counties by the 1850s. He argues that the patchy pattern of growth and decline in the various counties warn against applying any simple 'reaction to overpopulation' model to the study of late 19th century Scottish population history. First, although out-migration played a considerable role in depressing population growth rates during the latter 19th century, at the local level much population change was unrelated to migration. Some counties experienced population growth during periods of substantial out-migration. Similarly, prior to 1821, population loss through migration was common in many rural areas. Available data suggests migration was an integral component of rural Scottish society, even prior to the 1850s. Young adults, particularly males, regularly left their native lands in search of employment. Similarly, rural Scotland's nuptiality regime, characterized by considerably low levels and high proportions never marrying during the latter part of the 19th century, appears to have also been a long-standing characteristic of Scotland's rural society. The author concludes that it is a particular set of institutional elements, distinguishing Scotland from England, that underlie Scotland's rural demography during the period.

In their paper, 'Mortality and Migration in East Belgium During the Industrial Revolution: Town and Countryside Perspectives', George Alter and Michel Oris investigate the nature of the migration-mortality complex in East Belgium during the 19th century. Their analysis revealed that in-migration reduced death rates in the following year, suggesting that migrants exhibited lower mortality than urban natives. This finding indicates that in the short-term immigration increased the proportion of individuals with low mortality risks in the receiving urban centres. Further, it is in the largest industrial centres that migrant-native mortality differentials are greatest, suggesting that industrialization attracted a select population. They find, however, that this mortality differential, most pronounced in mid-century, disappears by 1866-1880. They ascribe this evolution to the changing migratory and industrialization phases of the century. Second, they test at the individual level whether indeed a selection process was behind the mortality differentials found. They find a strong relationship between family composition and decision to emigrate and an association between emigration and household mortality. Individuals from the healthiest households were the most likely to emigrate, suggesting migrants were indeed selected for factors associated with higher survival status. Finally, they explore whether this selectivity for greater survival extended to migrants' children and found that it did not, suggesting that the urban environmental characteristics of the period most likely eroded any preferential genetic or cultural characteristics migrants brought with them.

A few points dominated the discussion that followed. One, is the need to examine the counter factual. What would have happened to sending areas if there had not been rural-urban migration? Without an understanding of the 'what-might-have-been', we do not have a complete picture of the rural-urban migration process and its demographic and economic underpinnings. Another theme, is the need to ascertain what were the demographic effects of rural-urban migration in the sending areas. Did emigration increase fertility in rural sending areas? With regard to Livi-Bacci's call for comparative analyses, it was noted that a worthwhile model of differential country emigration should include an assessment of how internal migration was correlated with external emigration. Long and short-term fluctuations

In the first paper of the session, 'Economic Insecurity and Adult Mortality in Rural Sweden: 1760-1895', Tommy Bengtsson analyzes the relationship between population and economic change in rural Sweden. First, he examines whether, in the long-run and prior to the economic breakthroughs of the 19th century, Sweden could be characterized as a Malthusian society. His review of the association between price, wage and population change throughout the 17th and 18th centuries lead him to the conclusion that it cannot. For one, increases in wages followed the same pattern as increases in food prices. In addition, long-term decline in mortality levels began in the late 1700s, a full 75 years prior to the economic breakthroughs of the 19th century, leading to the conclusion that long-term mortality change was determined by non-economic factors. Second, he investigates whether in the short-term the living standards of pre-industrial Swedes were vulnerable to economic stress and whether such vulnerability led to changing demographic outcomes. An analysis of the relationship between changes in real wages and mortality, fertility and nuptiality at the aggregate level revealed that these demographic indicators were indeed sensitive to economic change. Particularly strong were the effects on fertility and adult mortality. In the final segment of the paper the author explores the causal mechanisms underlying these relationships and examines the effects of short-term economic stress at the micro-level in four parishes in Scania (a southern Swedish province). The most compelling findings are that the effects of economic stress on adult mortality outcomes were mediated by exposure to disease during the first year of life and that socio-economic status attenuated the effects of economic stress on mortality. These relationships were strongest during the earlier part of the 19th century.

In his paper 'Short-run and Secular Demographic Response to Fluctuations in the Standard of Living in England, 1540-1834' Roger Schofield examines whether and how Malthus's positive and preventive checks played a role in England's population growth during the 16th through 19th centuries. First, he examines the relationship between demographic and economic variations in the long-run. His findings suggest that the preventive rather than the positive check was involved and that its nature changed radically over time. For cohorts born before 1700 the nuptiality mechanism at play seemed to be proportions ever marrying, while for those born later in the 18th century it was marriage timing. The former period was distinguished by large proportions never marrying and the latter by first a fall and later a rise in age at first marriage. A feasible explanation lies in several changes in the economic structure. During the late 18th century these included the development of the 'protoindustry', changes in the demand for agricultural labour and changes in welfare policy, all conducive to more universal and earlier marriage. In the 19th century, a decline in the prevalence of the cottage industry and the return to a less generous welfare policy may have worked to raise age at first marriage. Second, the author explores whether England's escape from the positive check in the long-run held true in the short-run and whether the preventive check operated in the short-run as it had in the long-run. He reports that economic fluctuations exerted minimal influence on annual death rates, yet significant influence on nuptiality and fertility. Third, he explores the role of changes in standards of living and major economic transformations and finds that there was, over the three centuries, little variability in fertility response and some variation in nuptiality. The mortality response, however, appears to have changed considerably, suggesting that there was a time perhaps when the positive check was influential. The author concludes with an analysis of short-term infant and child mortality response to economic variations by age and sex and finds that there were gender differences. These results are a preview to upcoming analyses based on the Cambridge family reconstitution project.

A major issue highlighted in the discussions that followed these presentations concerns that of demographic regimes. Is it correct to compare Malthusian models across European countries without taking into consideration the demographic context in each country-time period? For instance, effects of economic crises on mortality surely must vary by levels of life expectancy. Likewise, the preventive check must work differently in contexts of high nuptiality regimes relative to those of low nuptiality regimes. In essence, it was argued that the effects of prices on demographic outcomes is not linear in all settings and at all times and that there are strong interactions that need to be examined. Similarly, variations in climate and harvest outcomes provide for important differences in production and its relation to price variations. Therefore, in order to achieve a truly comparative frame-work of the relationship between economic and demographic changes one needs to incorporate the demographic and economic context of a particular setting. Demographic regimes

In the first paper of the session, 'Malthus in Latin America: Demographic Responses During the XIXth and XXth Centuries', Alberto Palloni, Hector Perez Brignoli and Elizabeth Arias investigate whether Malthusian preventive and positive checks to population pressures were in evidence in the short-term in the Latin American context. First, they assess their role in ten Latin American countries during the modern period (1920-90). They find that in all the countries, representing diverse demographic trajectories, styles of economic development and political regimes, the patterns of demographic responses were as expected. Both marriages and births responded strongly and significantly to economic fluctuations, while the response of mortality was considerably limited. The patterns of response were similar to those that have been found for 19th century Western Europe. In addition, they found that although the positive check was considerably less visible than the preventive, fluctuations by cause of death were significant, at least after 1950. Second, they carry out a similar analysis for four cities and one country during Latin America's early post-colonial period (1820-1929). Both the positive and preventive checks were found to be at play. Births reacted according to expectation and more strongly than in either the European or modern Latin American contexts. Marriage responses were found to be similar across social and historical settings. And, although the mortality response was stronger in the early post-colonial period than in the other two settings, it exhibited a rather erratic pattern. The authors conclude that despite historical and institutional diversity, Malthusian short-term responses did exist in the Latin American setting, but believe that the identification of the preventive check is more robust than that of the positive check.

The paper, 'Malthusian Mythologies and Chinese Realities: from 200 Million to 1.2 Billion Chinese', by James Z. Lee and Feng Wang deals similarly with the identification of Malthusian preventive and positive checks. The authors propose that not only was Chinese population growth non-Malthusian, but that it differed in fundamental ways from the European case. First, they posit that despite the Malthusian implications of Chinese population growth since 1700, mortality rates did not increase accordingly. On the contrary, mortality levels either remained stable or declined as population grew. Second, they describe a socio-economic system characterized by increases in land and labour output and in consumption levels throughout the period. Third, they propose that China's escape from overpopulation was the result of two demographic-economic processes. At the societal level, a Boserupian process was at play, whereby population growth induced technological innovation. At the individual level, they propose a feedback loop, whereby economic circumstances induced individuals to modify their demographic behaviour. Individuals regulated their mortality through the practice of infanticide, their nuptiality by either marrying late or not marrying at all and, finally, they regulated their marital fertility through the practices of late starting, early stopping and wide spacing of births. These demographic behaviours were highly correlated with individual social status and economic circumstance. They conclude that given China's historical demographic system the definitions of the positive and preventive checks must be re-evaluated, or else would conclude they did not play a role in the Chinese case.

These papers generated discussions on issues similar to those highlighted in the previous session. Again, the relevance of demographic context with regard to proper assessments of economic-demographic variations and their effects on each other was highlighted. While the Palloni, et al. paper directly addressed this issue by comparing pre and post-demographic transition periods, it would have benefited by placing each 19th century city-period in a comparative demographic context. The Lee/Wang paper covers a very long period, yet does not clearly distinguish between pre and post-demographic transition period. Another important point concerns the reliability and representativeness of indicators and data. For example, can findings from studies of some villages be generalized to an entire country ? Likewise, how reliable are indicators used to ascertain economic conditions? Economic and demographic interactions

The first paper presented in this session, 'Malthus Revisited: Exploring Medium-Range interactions Between Economic and Demographic Forces in Historic Europe', by David S. Reher and Jose Antonio Ortega Osona, consists of an analysis of Malthusian population-economic interactions in the medium-term. The literature on Malthusian systems tends to report conflicting evidence concerning demographic-economic relationships in the long-term, but consistent and congruent evidence in the short-term. In the very short-term the direction of causality has been unquestioned: causality runs from economic conditions to demographic behaviour. On the other hand, it has been quite difficult to disentangle the causal mechanisms in the very long-term. This is partly due, the authors argue, to the circularity of Malthus's arguments, but also to the choice of indicators and time frames used. They propose that to ascertain the effects of population growth on economic trends a medium-term (25-300 years) perspective is more suitable. The effects of population size on economic conditions could only marginally be felt over the short run, while in the very long-run the links would have been obscured by exogenous long-range trends. To gauge the effects of economic constraints on demographic behaviour the short to medium-terms (10-100 years) are preferable. With regard to the positive check, the effects of fluctuations in wages and living standards on mortality should be present on a year to year basis. In terms of the preventive check, where marriage is a matter of individual choice, its sensitivity to economic conditions should also be visible in the fairly short term. The authors carry out analyses of these relationships based on economic and demographic series for England, North Italy and New Castile during the 16th through 19th centuries. They conclude that Malthusian links, in both directions, are more likely to be found in shorter cycles than in very long cycles. They found expected relationships in all three settings, although the Malthusian system worked somewhat differently in each. Further, exogenous factors played an important role in the short, medium and long runs.

In the second paper, 'Determinants of Mortality Variability in Historical Populations and Its Behavioural and Aggregate Consequences', Jose Antonio Ortega Osona examines the different aspects related to the variability of mortality at both the individual and aggregate levels. In the first two sections of the paper the author formalizes a dynamic model of personal mortality risk and identifies the conditions for the variability of food consumption and other factors that have an impact on mortality. In the third section he explores aggregate mortality experience in three historical populations: England, New Castile and North Italy (16th through 19th centuries). In the last section he examines the impact of excess mortality on the behaviour of people subject to recurrent mortality risk. He defines the positive relationship between the variability of factors determining mortality risk and average mortality levels as excess mortality and argues that excess mortality has implications for both individual behaviour and for the aggregate population. At the aggregate level, the stabilization of economic and environmental factors affecting mortality lead to an overall lower mortality rate and a higher population growth rate. Nonetheless, the process was not uniform across the three settings. In England, with low and stable mortality the effects were found to be negligible, but in North Italy and New Castile, which experienced unstable and high mortality, stabilization accounted for an important share of mortality gains throughout the period. Mortality stabilization was not explained entirely by economic factors; changes in exposure to disease may have also played an important role. Finally, he contends that a variety of behaviours, including food storage, can be explained in light of the implications of excess mortality at the individual level.

This session generated lively questions and discussions that transcended the issues brought up by the two papers presented. The issue of the applicability of historical price and wage data as measures of standards of living became a central topic of discussion. Are these measures appropriate in different settings? In the case of England, where a large portion of the population began working for wages at an early stage of its history, the indicators may be appropriate. But, what about in areas where say 10% of the population was in the wage sector and 90% in the subsistence sector; can one treat the average price as if everyone experienced it similarly ? Closing session

In the last session Robert W. Fogel presented his paper titled 'The Escape from Hunger and Premature Death 1700-2100: Europe, America and the Third World'. The paper is made up of four lectures that reflect twenty years of research concerning the secular decline in mortality. The essence of the lectures lies in the concept 'techno-physio evolution', which the author defines as a process whereby the synergism between technological and physiological improvements of the past three centuries, but particularly the current one, have led to a 'form of human evolution that is biological (but not genetic), rapid, culturally transmitted and not necessarily stable'. Advances in the nutritional and biomedical sciences have enabled researchers to better gauge the relationship between nutrition and physiological functioning. Indicators generated by such research suggest that the 20th century has been unique in terms of human biological evolution. During this century, human beings have increased their average body size by over 50%, average longevity by 100% as well as significantly improved their vital organs' robustness and capacity. Technological advances in food production and advances in economic, PUBLIC health and medical systems, which took over a century and a half to develop in Europe and North America, are now at the disposal of developing countries. The result has been an unprecedented pace in reduced mortality and improved health in developing regions. The process is an ongoing one. Developing countries are still working to catch up to current mortality levels found in developed countries, but similarly, improvements in longevity in developed countries have not yet run their course. Recent studies suggest that within the current state of medical technology simple changes in lifestyle, such as improved nutrition, the elimination of substance abuse and proper exercise regimens can increase life expectancy by 15 to 20 years. Future environmental changes and improvements in medical technologies may very likely raise the ceiling on life expectation. In effect, the author argues, the post-Malthusian world will be defined not by limits to the availability of natural resources but by the 'rate of technological progress, the supply of key scientific and technical skills, the preference for goods and services over leisure, the prevalence of the egalitarian ethic and the felicity of political systems'.

Elizabeth Arias

IUSSP, 3-5 rue Nicolas, F-75980 Paris cedex 20, France
Tel  +33 1 56 06 21 73 - Fax +33 1 56 06 22 04 - contact us