Infant and Child Mortality in Eighteenth Century France: A Function of Income?

Hajime Hadeishi
Bureau of Economics
Federal Trade Commission
Drop S-5016
Washington DC 20580
Telephone: 202-326-2761 Fax: 202-326-2625


I. Introduction and Overview

Empirical time series analyses of crisis mortality have shown that at aggregate levels, such as parishes or countries, variation in food prices explain little of the variation in crude death rates by the late eighteenth century. In contrast, Fogel suggests that even during the late eighteenth century, inequities in caloric intake forced many to live on the edge of subsistence. Small fluctuations in food supply would, therefore, have caused hardship for the poorer segments of the population. Further chronic malnutrition could affect mortality as well.

This paper quantifies the impact of grain price variation on infant and child mortality in Nuits-Saint-Georges. Nuits was a Burgundian town of relatively large size and broad economic activity. At any given time over 400 families could be found there with activities ranging from agriculture, crafts, trades, to high skilled services. Further, the tax rolls show that wide income and wealth variation existed in the town, even within occupational sectors.

Pre-Revolution and pre-fertility transition France was characterized by high levels of infant and child mortality. Similarly to the developing world today, water borne and communicable diseases contributed to large numbers of deaths in eighteenth century France. Further, in the poorest households family members and livestock frequently shared the same roof. Poor diets also contributed a large part to the susceptibility of pre-industrial populations to disease. Diets poor both in calories and vital nutrients adversely affected peasant survival.

Among the peasant population, infants and young children were the most likely to experience untimely deaths. Infants, in particular, suffered the most. They experienced the highest rates of mortality. Making things more difficult, in Old Regime France, birthrates were still very high. High birth rates contributed to higher infant death rates, especially at higher parities, for infants as well as their mothers. In brief, an eighteenth century French peasant child was generally born into an environment unfavorable to promoting long lives.

In this paper I determine which observable factors contributed to infant and child survival in Nuits. The problem is difficult because income affects both fertility and mortality. In a previous paper, I found that fertility is positively correlated with income. It is also likely that, all things being equal, higher fertility levels put higher birth order infants at risk. In practice, however, these effects might be mitigated by additional resources that a wealthier family could allocate towards these children. Further, the problem may be even more complex if the decision to have more children later in a marriage is a function of the mortality rate of the children born early in the union.

Although evidence from other studies on France suggests that by the late eighteenth century crisis mortality resulting from high grain prices explained less of the variation in crude death rates, food prices did indeed matter. My analysis shows that poor infants and children were more likely to die when grain prices rose. Moreover, other observable factors such as literacy and wealth levels clearly lowered the risk of mortality. In assessing how Europe's economic development and improved standards of living, we must examine the distribution of income, and hence calories.

II. The Historical Setting

My analysis focuses primarily on the sources of infant and child mortality at the household level. Nonetheless, this paper builds on the existing theoretical and empirical literature which has focused more on the macroeconomic relationships between economic conditions and demographic trends. My analysis stems from the classical economic tradition of Malthus (1793) and the recent cliometric work of Lee (1981), Weir (1984), and Fogel (1988). Malthus' Essay on the Principle of Population is best known for providing predictions of long run economic swings as a function of the population growth rate. However, Malthus' work also focused on the role of households in the decision making process and provides cross sectional implications. From Malthus' framework we can frame testable hypotheses of how mortality should vary across families at any given point in time. On the other hand, the more recent cliometric work of Lee, Weir, and Fogel focuses on long run time series relationships between economic indicators and demographic behavior. These works illustrate the long run relationships between economic trends and national or aggregated (at the city level) birth rates, death rates, and mortality rates.

Malthus is most famous for his dire prediction that the rate of population growth will exceed that of resource growth. Under this assumption, man's ability to multiply will be checked by periodic resource scarcity. In the extreme this leads to starvation or the positive check. To take this extreme scenario as the basis for analysis of all pre-industrial mortality is too narrow. It is obvious that the Malthusian positive check does not operate today (where the world's population is much greater than it ever was in the late eighteenth century when Malthus wrote the Essay). Moreover it is questionable whether the effect has been operational since the sixteenth century. Yet, the link between food supply and mortality in European history is still debated today.

Implicit in Malthus' work is the fact that many families lived on the brink of subsistence. The role of harvest crises on mortality has thus assumed primary importance in the debate on economic well-being and mortality. If this were the case then a bad local harvest could seriously affect the mortality rate. If markets were fully developed, as they are today in the industrialized world, a local grain harvest failure could be offset by food purchases from more distant markets or foreign trade. Trade would smooth out short run food shortages. But in pre-industrial economies with less developed markets, a local food shortage would force populations to cut into their grain reserves. This would then reduce the seed supply for the following year contributing to lower output. A key question is whether there were enough families that were so poor that short term fluctuations in the price of food would affect the overall mortality rate. I investigate this linkage at both at the aggregate and the household level.

The assumption that many in pre-industrial societies lived on the brink of starvation spawned a large literature. Most recently, Fogel (1988) presented estimated caloric intakes for late eighteenth century France and England. According to Fogel, in 1785 France the mean caloric consumption was 2,290 calories per consuming unit. However, for the lowest decile of the caloric intake distribution, the caloric consumption is only estimated to be 1,310 and 1,798 Kcals for the third decile. If one accepts Fogel's estimates of the distribution of caloric intake, large segments of the population suffered greatly because a poor harvest reduced their caloric intake even below their already meager rations. However, in the historical literature only a few studies have tried to test the cross-sectional implications of Fogel's estimates.

The first scientist to attempt to capture the impact of poor harvests on mortality, Meuvret (1951), argued that grain prices were a proxy for harvest conditions. In Europe, the majority of consumers were dependent on cereal products as their primary staple food. Poor harvests were caused by unfavorable climactic conditions. When harvests failed, prices would rise due to scarcity. Since peasant families spent a large fraction of their income on food, an increase in grain price would directly translate into a reduction of purchasing power. A harvest failure would seriously harm those living at the subsistence level. Therefore, if poor harvest are the primary source of mortality, a high grain price should result in a lower survival rate at the aggregate level. At the household level, a high grain price should increase the probability of an individual death. Pierre Goubert's analysis in Beauvais et le Beauvaisis actually found that a statistical correlation existed between grain prices and mortality crises in the seventeenth and early eighteenth centuries in France.

Recent cliometric work detailing the statistical impact of grain prices on mortality rates has greatly increased our knowledge of aggregate economic conditions and mortality rates. The seminal work is Lee (1981). Lee analyzed the relationship between 25 year centered moving averages of crude death rates and grain prices. He found that the period 1548-1640 displayed much higher correlation between the two series than the period 1746-1834. Similarly for France, Weir (1984) found that the sensitivity of mortality rates to grain prices fell over time. And across Europe, Galloway (1988) found that the overall trend is declining sensitivity of mortality to grain prices from the eighteenth to the nineteenth century. The implication for Meuvret's work is clear. At the beginning of the modern era, grain prices were strongly related to mortality rates. In Europe, the sixteenth century was very different from the nineteenth century where grain prices accounted for an insignificant portion of mortality. Yet, in France, mortality during the period 1740-1789 still displayed a great deal of sensitivity to harvest conditions (Weir, 1984). However, the relationship between 1740-1789 was actually negative. These results were truly contrary to expectation, but for the period 1790-1829 the effect reversed itself. Once again, higher grain prices tended to positively influence mortality.

Despite the declining importance of grain price in explaining pre-industrial European mortality, it is still likely that certain members of society suffered when grain prices rose. Malthus himself noted that people of differing income levels would suffer from higher food prices at differing rates. In 1793 he wrote:

"The labourer who earns eighteen pence a day and lives with some degree of comfort as a single man, will hesitate a little before he divides that pittance among four or five...[S]hould he have a large family, and any ill luck whatever, no degree of frugality, no possible exertion of his manual strength could preserve him from the heard-rending sensation of seeing his children starve." (Malthus, 91).

Understanding that the impact of grain price increases on mortality would differ by social class, Galloway (1981) separated the urban city of Rouen into rich and poor parishes from 1681-1744. He found that richer parishes demonstrated less response in mortality to increases in grain price. Yet within parishes, income had to vary across households. As a result a study at the household level would provide further evidence of how mortality varied across social and economic status. This paper attempts to address this.

III. Emprical Analysis and Conclusion

To construct my data set I combined marriage, baptism, and burial records to reconstitute individual family demographic histories for families married between 1744 and 1779. I next linked the demographic data to tax rolls to obtain family estimates of both income and wealth. The analysis relied on two methods. First I examined the correlation between infant mortality rates at the town level against grain prices. I found that a strong correlation existed between the two series. Second using a Weibull hazard analysis at the family level, I estimated the impact of a nine year centered moving average to capture the trend in grain prices along with a variable which captured grain price "shocks" (the difference between a year's grain price and the trend). I also controlled for a family's sector of occupation, whether the father was literate, dummy variables for two particularly high mortality years, and an average of three tax payments collected over fifteen years as a measure of family permanent income. The results of the hazard analysis revealed that children born to wealthier families survived grain price shocks significantly better than those in poorer families. Further, children born to literate families fared much better as well. And lastly, those born into the agricultural sector were significantly more at risk than those born into crafts, trades, and services. Household measures of income illustrate how mortality could differ across social and economic class. Using households as the unit of analysis in pre-industrial Europe will contribute greatly to our understanding of the decline of mortality. In late eighteenth century France, wealth therefore mattered for survival, and substantial portions of the population were still susceptible to food price shocks.


For a description of rural living conditions in pre-industrial England, see Laslett (1965).

2Poor nutritional levels would directly influence susceptibility to water borne diseases such as cholera or intestinal parasites (Livi-Bacci, 38).

3For an interesting look at whether modern population growth has affected world food prices see Sen (1994).

4 The argument that food shortages are due more to provisioning problems is not new. For a major works on the Irish potato famine and the role of provisioning, please see Mokyr (1983). For a new debate on the role of provisioning in England, see Fogel (1988) and an alternative evaluation of the role of the English government in Nielsson (1997).

5 The most directly related paper which tests mortality against income in the cross section is Weir (1995).

6 Although an individual should experience a higher mortality probability during a poor harvest, this by no means implies that all family members would be equally at risk. Age and birth order for young infants would matter. Further the distribution of food among household members may also be unegalitarian contributing to lower death probabilities for some family members over others.

7 Lee ran the centered moving average of crude death rates on the centered moving average of the current grain price, the grain price lagged one year, lag two years, lag three years, and lag four years. He found that the a 100% price increase in the grain price would increase the death rate by 22% from 1548-1640. Yet for 1746-1834 it would only increase it 15%.

8 In the same specification as Lee (1981), Weir found that a 100% increase in the grain price would contribute to a 50% increase in mortality from 1670-1739. For the period 1740-1789 the estimated effect is 45%, and for the period 1790-1829 the effect plummeted to 19%.

9 Notable studies on household mortality are Weir (1995) and Bengtsson (1996). Weir uses a cross sectional approach whereas Bengtsson models mortality in a hazard framework.

10 Results for different specifications will be produced upon request. Please contact me a copy of the paper.


Note: The references are for the entire paper and are a superset of articles cited in this conference summary.

Becker, Gary S. A Treatise on the Family. (Harvard, 1981).

Bengtsson, Tommy, "Combined Time Series and Life Event Analysis" in Reher, D and Roger Schofield, editors, Old and New Methods in Historical Demography (Oxford University Press, 1993).

Dupâquier, Jacques. Histoire de la Population Française aux XVIIe et XVIIIe Siècles (Presses Universitaires de France, 1979).

Dupâquier, Jacques. Histoire de la Population Française (Presses Universitaires de France, 1988). Four volumes.

Fogel, Robert W. "The Conquest of High Mortality and Hunger in Europe and America: Timing and Mechanisms" in Landes D. and H. Rosovsky, eds. Favorites of Fortune (Harvard, 1991).

Fogel, Robert W. "Economic Growth, Population Theory, and Physiology: The Bearing of Long-Term Processes on the Making of Economic Policy," The American Economic Review, 84.3 (June 1994), pp. 369-395.

Frankenberg, Elizabeth. "The Relationship Between Infant and Child Mortality and Subsequent Fertility in Indonesia: 1971-1991" (RAND mimeo, 1996).

Galloway, Patrick. "Short-Run Population Dynamics among the Rich and Poor in European Countries, Rural Jutland, and Urban Rouen," in Reher, D and Roger Schofield, editors, Old and New Methods in Historical Demography (Oxford University Press, 1993).

Galloway, Patrick. "Basic Patterns in Annual Variations in Fertility, Nuptiality, Mortality, and Prices in Pre-Industrial Europe," Population Studies, 42 (1988), pp. 275-303.

Labrousse, Ernest. Le Prix du Froment en France, 1726-1913 (S.E.V.P.E.N., 1970).

Laslett, Peter. The World We Have Lost, 2nd Edition (Methuen and Company, 1971).

Livi-Bacci, Massimo. Population and Nutrition: An Essay on European Demographic History (Cambridge Studies in Population, Economy, and Society in Past Time, 114, 1991).

Lee, Ronald. "Short-Term Variation: Vital Rates, Prices, and Weather" Chapter 9 in E.A. Wrigley and R.S. Schofield The Population History of England: 1541-1781, second edition (Cambridge University Press, 1989), pp. 356-401.

Malthus, Thomas. An Essay on the Principle of Population (Penguin Classics version of the original 1793 Essay, 1985).

Mokyr, Joel. Why Ireland Starved: A Quantitative and Analytical History of the Irish Economy, 1800-1850 (George Allen and Unwin, 1985).

Mroz, Thomas and David Weir. "Structural Change in Life-Cycle Fertility during the Fertility Transition," Population Studies, 44 (1990), pp. 61-87.

Nielsen, Randall. "Storage and English Government in Early Modern Grain Markets," The Journal of Economic History, 57 (March 1997), pp. 1-33.

Preston, Samuel H., editor. The Effects of Infant and Child Mortality on Fertility (Academic Press, 1978).

Saint Jacob, Pierre. Les Paysans de la Bourgogne du Nord au Dernier Siècle de l'Ancien Regime. (Publications de l'Université de Dijon, 1960).

Sen, Amartya. "Population and Reasoned Agency: Food Fertility and Economic Development," in Kiessling and Landberg (eds.) Population, Economic Development and the Environment (Oxford University Press, 1994), pp. 51-78.

STATA Corporation. Stata Statistical Software: Release 5.0 Reference Manual (Stata Corporation, 1997).

Thomas, Duncan, John Strauss, and Maria-Helena Henriques. "Child Survival, Height for Age and Household Characteristics in Brazil," Journal of Development Economics, 33 (1990), pp. 197-234.

van de Walle, Etienne and Helmut V. Muhsam. "Fatal Secrets and the French Fertility Transition," Population and Development Review, 21 (June 1995), pp. 261-279.

Weir, David R. "Life Under Pressure: France and England, 1670-1870," The Journal of Economic History, 44 (March 1984), pp. 34-65.

Weir, David R. "Family Income, Mortality, and Fertility on the Eve of the Demographic Transition: A Case Study of Rosny-sous-Bois," The Journal of Economic History, 55 (March 1995), pp. 1-26.

Wrigley, E.A. and R.S. Schofield. The Population History of England: 1541-1781, second edition (Cambridge University Press, 1989).