NOTES
1 Two laudable exceptions should be noted. The first is Witold Kula (1976: 119—31). In Kula's pioneering exercise, we see the first good sketch of the importance of differences in cost-of-living movements, though without any link-up with income measures and with
some limitations of commodity and geographic coverages.
Kula offered stylized ‘terms of trade' measures. His stylized nobility sold only rye, divided by a useful price series for a ‘basket of goods', consisting of good quality cloth, good quality paper, French wine, pepper, coffee, and sugar. Absent from this list are any meat, drink other than wine, or labour services, a luxury purchase we have emphasized. His stylized peasants sold a mixture of rye, oats, butter, and eggs in exchange for cloth, nails, and salt. The second exception is Williamson (1976), who explored movements in real income based on different cost-of-living indices for the different US income classes since 1820. Williamson's study, like the present, found that movements in relative prices typically accentuated the movements in nominal income inequality. The US relative-price effects he studied were, however, less dramatic than those revealed here. Besides these two exceptions, the closest approach to some of our present points was made by Wilhelm Abel (1973) but he mentioned only food versus other goods, and left the price effects unquantified.2.
‘The 1820s' here represent a compromise post-war benchmark. In some cases, the phrase means the initial postwar era of recovery on the Continent and agricultural depression in England, 1815—20. It also serves as a link to Angus Maddison's (2001) estimates for 1820 in his study of the world economy and to the argument by O'Rourke and Williamson (2000) that international price convergence may not have progressed between the 16th century and the 1820s.
In other cases, it refers to the 1831 output benchmark date for the United Kingdom or to the Morrisson and Snyder (2000) benchmark estimates of French inequality in 1836. And for some continental countries the disruption of the French war era 1790—1815 will require stopping the evidence at 1790 instead of the 1820s. Similarly, we will at times equate the 1640s or even the early seventeenth century with 1650 and the 1740s with 1750 as turning points. The dating of turning points has to be flexible, since different series changed trends at different times.3.
The sensitivity of mortality to food scarcity is confirmed by several studies in this volume.
4.
So say not only Table 6.1's estimates for Swedish females 1750—9 but also the reconstitution experiments of Bengtsson and Oeppen on population data from Scandia in southern Sweden 1650—1750 (Bengtsson and Oeppen 1993). Apart from mortality crises around 1675 and 1705, Scandia's mortality seemed no worse than English over this century.
5.
This way of summarizing the inequality of life spans is the conventional way which tends to focus on class and to view the sur vival of infants as something experienced by their parents' class. A more logical approach would be to view all persons as separate individuals, and to follow the inequality in life spans and in consumption per lifetime. On this more logical approach, inequality of life spans in western Europe has been dropping ever since 1750, thanks mainly to the elimination of infant mortality. It has not been dropping for the world as a whole, however (Bourguignon and Morrisson 2002).
6.
Table 6.1's distinctions between the full-rent and no-rent expenditures for the upper classes will be discussed in Section 6.
7.
For the strong conclusion that workers' abilities to purchase food were declining in most European settings across the second half of the eighteenth century or longer, see Phelps Brown and Hopkins (1957, 1959, 1981); Braudel and Spooner (1966); van Zanden (1999); and Allen (2001, 2003).
8.
The monotonic rise of global income inequality since 1820 is documented by Bourguignon and Morrisson (2002).
’■ This is not to imply that all new goods were luxuries. The New World introduced the potato and other new foods that were to become staples. Yet, the introduction of these was not sufficient in its impact to prevent the rise in the real price of staple foods relative to luxuries before 1820. Two other likely exceptions were health related: cotton clothing and other aids to sanitation might have helped the poor as much as the middle- and upper-income groups. Even these new goods had a not-so-egalitarian side, to the extent that they raised the labour supply, bidding wages down and rents up.
10 For example, see Feenstra (1994, 1995) and Hausman (2003).
11' For example, in 1784—6, spice imports were less than 1% of the value of imports into Great Britain (Davis 1979: 110). They would have taken an even smaller share of income in the top 20% of families.
12' That labour was increasingly cheap relative to most products and the overall cost of living is evident in Table 6.4 and in the real-wage studies of van Zanden (1999), Allen (2001, 2003), and earlier authors. The cheapening of capital inputs follows mainly from the drop in interest rates (Allen 1988; Homer and Sylla 1991; Clark 1998).
13' On a musical note, another luxury that became cheaper relative to bread or the overall cost of living was a London opera ticket. While other nominal prices rose between the 1720s and 1786—1878, opera tickets stayed fixed at 10s 6d for boxes and the pit, and at 3—5s for the gallery (Hunter 2000: 35).
14' For example, see Lindert and Williamson (1982, 1983) and van Zanden (1995).
15' This statement uses the social tables in Lindert and Williamson (1982) and assumes that home production could have been 100% of income for lesser freeholders and husbandmen, and 10% for labourers, vagrants, paupers, and cottagers.
16 On housing rents in Holland from the mid-sixteenth century on, see van Zanden (2000) and the sources cited there. A series for Bruges from 1500 on is given by Verlinden (1959—65). Clark (2002) presents new series for England and Wales 1550-1909.
17' For an application of the unit-elastic assumption about housing to the economic history of inequality, see Williamson (1985).
18 Our more detailed treatment in the working paper version (Hoffman et al. 2000: Tables 4-6) sets extreme bounds on the importance of this residence-ownership effect. As warned in the text here, the bounded estimates that give maximum impact to this effect suggest that it may have heightened the rise of real inequality in rising-rent periods like the sixteenth century.
”■ This chapter confines itself to sketching movements in early modern inequality in income before, rather than after, taxes and transfers. The pre-tax, pre-transfer focus avoids the difficult task of exploring fiscal incidence, a topic too vast to undertake here. So for the time being, we must perpetuate an inconsistency shared by most of the literature presenting real-income time series. We use a price deflator appropriate to disposable income—that is, after-tax and after-transfer income—as a denominator for original income, from which taxes and transfers have not been netted out.