دستمزدهای واقعی و خانواده : تنظیم دستمزدهای واقعی برای تغییر جمعیت شناسی در انگلستان قبل از مدرن

This paper uses demographic data drawn from Wrigley et al.'s (1997) family reconstitutions of 26 English parishes to adjust Allen's (2001) real wages to the changing demography of early modern England. Using parity progression ratios (a fertility measure) and age specific mortality for children and parents, model families are predicted in two reference periods 1650–1700 and 1750–1800. These models yield two levels of interesting results. At the individual family level, we can measure how different families' real wages changed over the family life cycle as additional children were born. At the aggregate level, we can predict thousands of families using Monte Carlo simulation, creating a realistic distribution of median family real wages in the economy. There are two main findings. First, pregnancy and lactation do not create cyclical effects in the family's income. Instead, most families' welfare ratios decline steadily across the family life cycle until children begin to leave the household, increasing the welfare ratios. Second, Allen's real wages understate or match the median of the predicted demography-adjusted distributions.

The flurry of scholarship in the last decade reconstructing historical real wages around the world has fundamentally changed the way economic historians understand the Great Divergence debate. By comparing nearly identical consumer price baskets and labourers' wages, first Allen and then his collaborators were able to construct comparable real wages for most parts of the world including North America, Latin America, Europe, Africa, India, and China (Allen, 2001, Allen et al., 2011 and Allen et al., forthcoming). These papers have generally refuted the California school position that parts of China and India were just as developed as Europe in the eighteenth century. Real wages were quite low around the world from the sixteenth century onwards with notable exceptions in Britain and the Netherlands. This scholarship has been tremendously helpful in reconstructing the economic history of the world, but in order to make international comparisons, Allen and his co-authors made certain assumptions about family size and composition over time and across countries. Allen based his family size assumptions on an English gardener from Ealing, who had a wife and four small children, described by Sir Frederick Eden in The State of the Poor in the 1790s. Therefore, Allen assumed that the average family size and composition around the world was similar and consisted of the equivalent energy needs of three adult males (Allen, 2009, p. 29). Humphries has recently criticized Allen's constant family size as being too low for England, thus making his English real wage estimates too high. There is therefore room in the literature for a careful reworking of the relationship between family size and real wages. In general Allen's family size assumption may be justifiable because of the paucity of detailed demographic information that would allow historians to precisely vary family size over time for many countries. Family size is also difficult to proxy with other demographic variables because it depends not only on fertility measures but also on the mortality of children. However, there is good demographic data for England from which we can attempt to understand how changing family size influenced real wages over time. This paper will use demographic data drawn from Wrigley et al.'s (1997) family reconstitutions of 26 English parishes to adjust Allen's real wages to the changing demography of early modern England. Using parity progression ratios (a fertility measure explained later) and age specific mortality for children and parents, model families are predicted for two reference periods, 1650–1700 and 1750–1800. We can then study how the changing size and composition of an individual family affected its ‘welfare ratio’1 over the family life cycle. The welfare ratio is the wage earned by the father divided by the consumption requirements of the family in any given year. In addition, Monte Carlo simulation can be employed to predict thousands of families in each reference period, providing a realistic distribution of welfare ratios based on the different families possible. These distributions can then be compared with Allen's original figures to measure the influence of changing family size and structure on real wages.

This paper has presented demography-adjusted welfare ratios for two broad reference periods, 1650–1700 and 1750–1800, showing that demography-adjusted welfare ratios are either higher than or match Allen's original real wage estimates. These results are robust to the inclusion of remarriage in the model, and any women and children's labour force participation would only increase the demography-adjusted figures further. The median family size in terms of consuming units was 2.69 in 1650–1700 and 2.98 in 1750–1800. Family size was much smaller than previous historians have argued for two reasons. First, relatively wide birth spacing ensured that not all children were present in the household at the same time and that the children each reached their peak consumption at different times. Second, mortality of children and adults significantly reduced the number of children born per family and the mouths that needed to be fed by the family. These factors kept family size at a relatively low level. However, Allen's original real wage series stretch from 1264 to 1914. While these broader time periods cannot be incorporated into the model, it is possible to speculate about how changing fertility and mortality might have affected the distribution from the sixteenth century onwards. Fertility, measured by the net reproduction rate (NRR), was higher from 1550–1650 than from 1650–1700, and infant and childhood mortality was lower (Wrigley et al., 1997, pp. 239, 250–51, 290, 614). This suggests that family sizes were larger in the period 1550–1650 than in 1650–1700, but not quite as large as in 1750–1800 because fertility was not as high. Thus, Allen's real wages are probably close to the demography adjusted figures until the mid-seventeenth century when decreasing fertility and increasing mortality made family sizes smaller and drove the demography-adjusted welfare ratio above Allen's figure. By the mid-eighteenth century, however, family sizes had again grown because of increasing fertility and declining mortality of children and adults. Allen's real wage was probably again approximately equal to the demography-adjusted welfare ratio. Family sizes were likely at their largest during the first half of the nineteenth century because fertility was unprecedentedly high and mortality was low. Thus, it is possible that the demography-adjusted welfare ratio could have fallen below Allen's real wage figure. In the second half of the nineteenth century, fertility fell to levels similar to those in 1650–1700, but childhood mortality rates also fell substantially (Woods, 2000, pp. 6, 253). The combined effects of these two processes likely kept family sizes at levels similar to those in 1750–1800, suggesting that Allen's real wages were close to the demography-adjusted welfare ratios. It is also important to note that the scale of the adjustment for the demographic factors is really quite small. The largest gap between the demography adjusted welfare ratios and Allen's real wages was the difference in the period 1650–1700, but it only represented an 11.77% increase in the welfare ratio. In order to halve the median building labourer welfare ratio in 1750–1800 to 1.225, the median family size would have to equal 6.02 consuming units. Families only reached this level of consumption in 2.55% of the total family years lived in 1750–1800. Thus, although adjusting for demography might raise or lower Allen's figures slightly, the demography adjustment is unlikely to change the general trend of the real wages Allen produced. Therefore, scholars wanting to improve upon Allen's method might find querying some of his other assumptions, for instance the constant 250-day work year and the reliance on male wages rather than household income, a more productive way of moving forward.