Land and inequality in Canada 1870–1930

Abstract

The great world economic and globalisation boom of the pre-First World War era was accompanied by great inequality in the distribution of income and wealth particularly during industrialisation, with the new world European settler economies being no exception. Canadian wealth inequality over the period 1870–1930 was also substantial and is examined using probated estates from the Eastern Judicial District of the province of Manitoba and Wentworth County, Ontario. However, wealth inequality is found to be less pronounced in frontier Manitoba relative to Ontario with higher and more dispersed rates of land ownership in the West as well as lower wealth levels and greater farm employment, as the key factors in this difference. This suggests that the farm economy of pre-First World War Canada was associated with greater equality of wealth. One of the inevitable effects of Canadian industrialisation and economic development was a rise in wealth inequality but the process of western settlement and associated free grants helped mitigate it. By extension, global economic inequality might also have been mitigated during this period by the presence of agricultural frontiers with subsidised land grants.

Keywords:

• wealth distribution
• probate records
• settlement
• North America

JEL Classification:

• N5
• N9
• D3

Acknowledgements

The author gratefully acknowledges the financial assistance of the Social Sciences and Humanities Research Council of Canada. As well, this paper benefited from the helpful comments of participants at the Historical Inequality and Mobility: New Perspectives in the Digital Era, University of Guelph, May 25–27, 2012 as well as the comments provided by the reviewers of this journal.

Notes

¹See, for example, Kirk, Promise (1978), Thernstrom, Bostonians (1973), Katz, People (1975). Darroch/Soltow, Property (1994).

²See Ohlsson et al., ‘Changes’ (2008), Piketty et al., ‘Wealth Concentration’ (2006), Lindert, ‘Three Centuries’ (2000).

³Maddison, ‘Economic Performance’ (2005).

⁴This period was not without discontinuities or changes in the rate of growth. See Green/Sparks, ‘Population’ (1999) and Inwood/Stengos, ‘Economic Growth’ (1991). As well, there were regional differences in output and output growth. See Inwood/Irwin, ‘Land’ (2002).

⁵Data Sources: EH.Net, International Historical Statistics; Mitchell, International (1992, 1993) and Green/Urquhart, ‘New Estimates’ (1987).

⁶Inequality characterized most North American jurisdictions. Between 1871 and 1899, Darroch finds the Gini coefficient in Toronto ranging from 0.690 to 0.598 while in Massachusetts over the 1879 to 1891 period it ranged from 0.80 to 0.83. For a summary table of comparisons, see Di Matteo/George, ‘Canadian Wealth’ (1992).

⁷Davies/Shorrocks, ‘Distribution’ (2000), 607. Gini coefficients for income in developed countries currently range in the 0.3–0.4 while for wealth the range is 0.5–0.9.

⁸For Canada: Siddiq, ‘Size Distribution’ (1988), Osberg/Siddiq, ‘Inequality’ (1988), Osberg/Siddiq, ‘Wealth’ (1993), Darroch, ‘Industrialization’ (1983), Siddiq/Gwyn, ‘Importance’ (1991), Di Matteo/George, ‘Canadian Wealth’ (1992), Di Matteo/George, ‘Patterns’ (1998), Gwyn/Siddiq, ‘Wealth Distribution’ (1992), Darroch/Soltow, Property (1994), Bouchard, ‘Economic Inequalities’ (1998), Baskerville, ‘Women’ (1999). For the USA: Main, ‘Probate’ (1975), Jones, Wealth (1980), Burchell, ‘Opportunity’ (1987), Bolton, ‘Inequality’ (1982), Soltow, Men (1975), Soltow, ‘Inequality’ (1979), Atack/Bateman, ‘Egalitarianism’ (1981), Newell, ‘Inheritance’ (1986), Newell, ‘Wealth’ (1980), Herscovici, ‘Distribution’ (1993), Pope, ‘Households’ (1989), Gregson, ‘Wealth’ (1996), Ferrie, ‘Wealth’ (1994), Steckel, ‘Poverty’ (1990), Steckel/Moehling, ‘Rising Inequality’ (2001), Stewart, ‘Migration’ (2006), Clay/Jones, ‘Riches’ (2008), Walker, ‘Opportunity’ (2000), Canaday, ‘Property’ (2008), Stewart, ‘Economic Opportunity’ (2009).

⁹Shanahan, ‘Distribution’ (1995) and Galt, ‘Wealth’ (1985) and McAloon, Idle Rich (2002). See also Rubinstein, ‘Distribution’ (1979). For a reference on the use of probate records in English economic history, see Owens et al., ‘Measure’ (2006).

¹⁰Ohlsson et al., ‘Changes’ (2008).

¹¹Kuznets, ‘Economic Growth’ (1955) and Modern Economic Growth (1966).

¹²According to Jeffrey Williamson ‘increasing regional inequality is generated during the early development stages, while mature growth has produced regional convergence or a reduction in differentials’ Williamson, ‘Regional Inequality’ (1965), 44.

¹³Williamson/Lindert, American Inequality (1980), Lindert/Williamson, ‘Growth’ (1985), Lindert ‘Comparative History’ (1991).

¹⁴Higgins/Williamson, ‘Explaining Inequality’ (1999), 8–9. Other studies in this category are Chenery et al., Redistribution (1974), Ahluwalia, ‘Inequality’ (1976), Anand/Kanbur, ‘Inequality’ (1993), Deininger/Squire, ‘New Data’ (1996) and Deininger/Squire, ‘New Ways’ (1998).

¹⁵Williamson, ‘Growth’ (1998) and ‘Globalization’ (1996). Higgins/Williamson, ‘Explaining Inequality’ (1999). The strict Kuznets curve asserts that demand side forces emanating from technological and structural economic change favour the demand for capital and skills which first worsens inequality and then is moderated during a catch-up phase. Globalisation and economic openness leads to trade which leads to changes in factor demands and incomes, which also can affect inequality. Finally, the cohort effect finds that large mature earnings age cohorts (as opposed to young or old cohorts) are associated with lower inequality.

¹⁶Green, ‘Regional’ (1967) and Green, Regional (1971) and Green, ‘Regional Inequality’ (1968/1969).

¹⁷Inwood/Irwin, ‘Green Hypothesis’ (1998), 1. According to Green, ‘Regional’ (1967) and Green, Regional (1971), in Canada, the period 1890–1910 saw a widening of regional income inequality with the disparity widening further from 1910 to 1929 and then some narrowing of the disparity between 1929 and 1956. Inwood/Irwin, ‘Green Hypothesis’ (1998), 23, calculate descriptive income inequality measures for 1870 and 1890. Using districts weighted by population, they find that the value of the Gini coefficient in Canada was 0.14 in 1870 and 0.15 in 1890. As well, the income shares of the wealthiest and poorest fifth of the population remained about the same at 27 and 13%, respectively. See also Inwood/Irwin, ‘Land’ (2002).

¹⁸Inwood/Irwin, ‘Green Hypothesis’ (1998), 18.

¹⁹In related work, Beaulieu/Emery, ‘Economics’ (1998), advanced the hypothesis that the linkages from the Prairie wheat economy were probably not very significant to Ontario relative to the importance of the Ontario economy itself. In other words, Ontario and the Prairies may not have been as interdependent as the staples theory requires and as long as markets are not well integrated, one would expect continued divergence rather than convergence in regional wealth and income even in the presence of robust growth.

²⁰Di Matteo, ‘Patterns and Determinants’ (2001).

²¹Williamson, ‘Growth’ (1998), 259. For another discussion of age and inequality also see Davies, ‘Wealth’ (1996).

²²Green/Sparks, ‘Population Growth’ (1999).

²³Davies/Shorrocks, ‘Distribution’ (1999) 643–44.

²⁴Skott, ‘Increasing Inequality’ (2011).

²⁵For example, the Royal Commission on the Distribution of Income and Wealth (1979) for the UK found the reduction in the wealth share of the top tiers of the wealth distribution over the 1972–1976 period to be the result of a rise in land prices and a drop in the stock market.

²⁶Keay, ‘Engine’ (2007).

²⁷In a sense nothing is ever free as there is opportunity cost between taking a piece of free grant land and the next economic opportunity available to a farm family.

²⁸Norrie, ‘Rate of Settlement’ (1975), 411.

²⁹Lucas, ‘Land’ (1966). Small farming did not thrive and the importance of wool emphasised large estates.

³⁰Norrie et al., A History (2008), 96.

³¹Lewis, ‘Farm’ (2001), 175.

³²Norrie, ‘Rate of Settlement’ (1975) 410.

³³Pomfret, Economic (1993), 183. See also Martin, ‘Dominion Lands’ (1973).

³⁴The type of farming is also related to whether agriculture results in family size or estate size farms. For example, sheep farming in New Zealand resulted in more estate sized farms.

³⁵Studness, ‘Economic Opportunity’ (1964).

³⁶Wentworth County was one of 41 southern Ontario counties in the nineteenth century each with a probate court. Manitoba was divided into judicial districts for probate which by 1919 totalled six: Eastern, Southern, Western, North, Central, Dauphin.

³⁷The boom period seemed to spark more extensive rather than intensive growth in the Hamilton–Wentworth region as per capita wealth did not rise as dramatically during the boom as in more western regions of Canada. See Di Matteo, ‘Booming Sector’ (1993) and Di Matteo, ‘Boom’ (2004).

³⁸East and West Flamborough, Beverly, Ancaster, Glanford, Binbrook and Saltfleet.

³⁹The Loyalists were British subjects who left the USA after the Revolution.

⁴⁰Statutes of Canada, 22 Vict., Cap. 93, 1858.

⁴¹Howell, Law (1880), 155.

⁴²For information on the collection of the original Wentworth County Data set of 405 census-linked observations and criteria and approach to linkage, see Di Matteo/George, ‘Canadian Wealth’ (1992). An additional 39 decedents for 1902 were census-linked later bring the data up to 444. For 1872 and 1882, probate only provides personal estate and therefore the real estate data for these original census-linked estates was obtained from assessment rolls and therefore no additions were made to the data for these years. However, had these estates been utilized there would be a total of 72 for 1872 and 115 for 1882.

⁴³Approximately 37 of the individuals in the 1927 data actually had an application date for probate in 1926 but their files were located within the date sequence for 1927.

⁴⁴The decedents were organized in alphabetical order by year probated and an effort made to trace the proportion of probated decedents in each year in the census using Ancestry.com. As the number of decedents expands with each additional year cross-section, a smaller proportion was selected for census tracing. The proportions were 60% of decedents in 1907, 50% in 1912, 45% for 1917 and 30% for 1922. No estates from 1927 were selected for linkage given the distance from the 1911 census and lower success rates for 1917 and 1922 compared to 1912.

⁴⁵For the period 1907–1922, a total of 352 census-linked probated decedents were constructed which when added to the 444 already available bring the total to 796.

⁴⁶Census of Canada.

⁴⁷Norrie et al., A History (2008), 193–211. See also: Morton, Manitoba (1957); Artibise, Gateway City (1979).

⁴⁸According to Howell, Law (1880), 325–26:

The inventory should contain a statement of all the goods, chattels, wares and merchandize, as well moveable as not moveable, which were of the person deceased at the time of his death within the jurisdiction of the court. A proper inventory should enumerate every item of which the personal estate consisted, and should specify the value of each particular. But unless by order of court, or in obedience to a citation, an inventory does not set forth the goods and chattels in detail.

Probate instructions do not specify how asset value was assigned. For real estate, livestock and personal property the evidence suggests that it was market value. Sometimes, property was sold and its selling price recorded in the inventory, whereas more often it was an estimate of what the property would fetch if sold. Financial assets by their nature were precisely recorded. Mortgages held, the amount of insurance payments, and bank account balances were precise amounts. In addition, real estate was usually recorded net of any mortgages outstanding.

⁴⁹The inventory categories were:(1) household goods and furniture, (2) farm implements, (3) stock in trade, (4) horses, (5) cattle, (6) sheep and swine, (7) book debts and promissory notes, (8) moneys secured by mortgage, (9) life insurance, (10) bank stocks and other shares, (11) securities, (12) cash on hand, (13) cash in bank (14) farm produce, (15) real estate, (16) other personal property. For further discussion of probate records used in the Ontario data, see Di Matteo, ‘Determinants’ (1997), Di Matteo, ‘Wealth’ (1998), Di Matteo/George, ‘Canadian Wealth’ (1992) and Di Matteo/George, ‘Patterns’ (1998). For an evaluation of probate sources, see Elliott, ‘Sources’ (1985), Osborne, ‘Wills’ (1980) and Wagg, ‘Bias’ (1990).

⁵⁰For example, in Wentworth County Ontario, the proportion of adult deaths with estates probated rose from 23 to 36% between 1872 and 1892. See Di Matteo, ‘Wealth Holding’ (1990), 46.

⁵¹Information on cause of death was not available in the probate records.

⁵²In Manitoba, the provincial Succession Duty Act did not apply: (1) To any estate the value of which after payment of all debts and expenses of administration, does not exceed 4,000 dollars per year, nor (2) To property passing under a will, intestacy or otherwise, to or for the use of the father, mother, husband, wife, child, grandchild, daughter-in-law or son-in-law of the deceased (…) where the value of the property so passing does not exceed 25,000 dollars in value (…) These exemptions where somewhat less generous than Ontario where its Succession Duty Act did not apply :(1) To any estate the value of which, after payment of all debts and expenses of administration, does not exceed $10,000; nor (2) To property given devised or bequeathed for religious, charitable or educational purposes; nor (3) To property passing under a will, intestacy or otherwise, to or for the use of the father, mother, husband, wife, child, grandchild, daughter-in-law or son-in-law of the deceased, where the aggregate value of the property of the deceased does not exceed $100,000 in value.

⁵³Average wealth higher in the east relative to the west does fit in with Green's work that regional income inequality increased in Canada overall because of the expansion of low-income agriculture in the west and high income manufacturing in the east.

⁵⁴Green/Urquhart, ‘New Estimates’ (1987).

⁵⁵LOWESS or Locally weighted scatter plot smoothing is a simple non-parametric data smoothing technique that is less sensitive to outliers. See Cleveland, ‘Regression’ (1979), Cleveland, The Elements (1985), and Cleveland, Visualizing Data (1993). STATA 11.0 was the estimation package used for these figures.

⁵⁶While total wealth was increasing in the Hamilton–Wentworth area it appears the number of decedents was increasing faster than wealth. Moreover, the proportion of female wealth holders was growing and women's wealth while rising was generally lower than men's.

⁵⁷For a discussion of various inequality measures, see Cowell, Measuring (1977).

⁵⁸Source: Tynan, ‘Patterns’ (1971/1972), Table 2.

⁵⁹LOWESS profiles of the inequality measures of wealth versus year find a generally upward sloping profile with a decline in inequality after 1920. This could be interpreted as evidence of a Kuznets type relationship.

⁶⁰See Urquhart, National Product (1993).

⁶¹For example, Ontario's rural population share declined from 78% in 1871 to 39% by 1931 and was accompanied by an increase in urbanization. The proportion of Ontario men gainfully employed in agriculture declined from 52.2% in 1891 to 27.2% by 1931 Source: Census of Canada. Another feature of the post-Confederation era was expansion of Canadian savings rates [Green/Urquhart, ‘New Estimates’ (1987)] and growth of the financial sector [Neufeld, Financial (1972)].