ثروت و تقاضا برای بیمه عمر: شواهد از انتاریو، 1892

Our analysis of the life insurance holdings of male probated decedents in Ontario in 1892 demonstrates a negative correlation between the level of personal wealth and the demand for life insurance. Consistent with the theoretical literature on the demand for life insurance, and counter to the findings in much of the empirical literature, self-insurance was a substitute for market purchases of life insurance where self-insurance capabilities are a by-product of wealth accumulation. Our evidence suggests that households primarily demanded life insurance when they lacked accumulated reserves, or wealth, for self-insurance, often early in the life cycle. The growth of the life insurance industry reflected the growth of urban population relative to farming and an increased dependence on the head of household’s earnings

Over the course of economic development in Canada and the US, rising levels of investment in human capital through work experience and education have driven much of the rise of real incomes.1 While the reliance on human capital investment resulted in higher long-term levels of consumption for North American households, it was also a riskier consumption stream as households became more reliant on a smaller number of more educated workers.2 Unlike bonds or other financial savings instruments, human capital is an asset that pays returns only when the individual in whom it is vested is alive. For households that relied on a tangible asset like land to generate income, the death of the household head, while tragic, was not as important an event for the surviving dependents’ consumption as it was for surviving dependents in households that relied on the household head’s human capital to generate household income. As Horrell and Oxley (2000) argue, despite what we know about the range of risk mitigation strategies available to households, we do not know very much about how households chose to address the increase in risk that accompanied the increase in reliance on a household head’s human capital. For the purpose of insuring a household’s investment in human capital, was the purchase of market insurance coverage the preferred (or primary) strategy or was self-insurance through precautionary savings or accumulated wealth? The answer to this question is of considerable importance if we are to understand the development of insurance markets. For example, a common theme in historical studies of life insurance is that life insurance was an unsought asset due to the ignorance or improvidence of consumers, or due to consumers perceiving the insuring of a life as a suspect or immoral contract. As such, the rise of the life insurance industry and extent of insurance coverage was largely a function of the aggressive sales efforts of life insurance salesmen.3 Another common theme in the history of life insurance emphasizes the importance of problems of asymmetric information between buyers and sellers of life insurance precluding households from purchasing coverage at affordable rates. Thus, information and social barriers and market imperfections frustrate the demand for life insurance. The growth of life insurance markets is largely explained by changing preferences, and the reduction, if not elimination, of market imperfections and can be interpreted as representing an increase in the level of insurance coverage for households. A less explored factor in the development of life insurance markets is the availability of substitute strategies and assets for the market purchase of a life insurance contract. From this perspective, the development of life insurance markets would reflect favorable changes in the relative price of life insurance contracts versus its alternatives.4 For example, favorable income tax treatment of life insurance premiums provided through the workplace, and/or employers subsidizing the insurance premia for employees, is an obvious explanation for the expansion of life insurance coverage after World War II. In contrast to the preceding perspectives on the growth of life insurance markets, extensions of market life insurance coverage reflect changes in the compositions of household insurance portfolios rather than an increase in a household’s level of insurance coverage. The relative importance of market insurance purchases and self-insurance is also important for understanding the impact of government insurance schemes on private behavior. Do government programs such as Social Security or Worker’s Compensation Laws reduce precautionary savings or do they replace purchases of comparable commercially provided products such as life insurance? (Feldstein, 1974; Kantor and Fishback, 1996; Lewis, 1989). If household portfolio choices are interdependent, then these questions do not have obvious answers as changes in the portfolio due to government policy would reflect the direct effects of policy and the indirect effects related to the degree of substitutability of assets in the portfolio (Mayers and Smith, 1983). Finally, an understanding of a household’s demand for life insurance over the life cycle can shed light on the demographic forces that were behind the rapid growth of Canadian financial markets in the late 19th century as Canada went from being a relatively young frontier society to an older, more settled economy. The aggregate saving ratio rose from an average of 8.7% of GDP in the 1870s to 15.7% during the first decade of the 20th century.5 In real terms (1900 dollars), between 1870 and 1910, the value of financial assets in Canada rose from 119 million to 1.5 billion dollars. Growth was especially pronounced in the assets of chartered banks, mortgage and loan companies, trust companies, and life insurance companies (Neufeld, 1972). In particular, Canadian life insurance companies expanded dramatically in the face of competition from British and American life insurance firms and became important participants in Canadian capital markets, with investments in real estate, bonds and mortgages (Drummond, 1962). The nominal value of life insurance in force in Canada rose from 43 million dollars in 1870 to 856 million dollars by 1910. The total assets of Canadian life insurance companies rose from a value of 21 million dollars in 1890 to 171 million by 1910.6 Life insurance in Canada was indeed an important financial asset. We seek to contribute to the understanding of the role of life insurance in household portfolios by taking advantage of a micro-data set with information on wealth holdings of Ontario probated decedents in 1892 (Di Matteo, 1996, Di Matteo, 1997 and Di Matteo, 1998). The information on wealth includes the value of household life insurance paid at death. In addition, these records have been linked to 1891 Census manuscript data that provides us with details on household characteristics. As few historical studies of insurance demand exploit micro-data our study contributes information that to date has been scarce (Kantor and Fishback, 1996). Nineteenth century data can potentially provide clearer insights into the demand for life insurance over the life cycle in the presence of self-insurance alternatives relative to data from the post-World War II era. First, the life insurance industry in Canada and the US was well developed by 1892 (Bliss, 1990; Drummond, 1962; Zelizer, 1979). Canadian households were also capable of accumulating total wealth at annual rates ranging from 1.7 to 9.7% particularly during earlier portions of the life cycle (Di Matteo, 1998). Thus, by the late 19th century, market insurance and self-insurance were both available to consumers. Second, with data from 1892, we have less of a problem controlling for some pertinent but unobserved household characteristics than studies using more contemporary data sources. For example, workers were unlikely to have life insurance provided by their employers as a non-wage benefit than were workers after World War II (Pedoe and Jack, 1978). Therefore, we are confident that we are observing voluntary purchases of insurance by individuals. Similarly, Canadian households in 1892 did not expect to benefit from government pension arrangements with survivor’s benefits or Worker’s Compensation benefits. Our empirical evidence supports the proposition that for all but the wealthiest of men, households with higher levels of wealth purchased less life insurance. While this behavior is consistent with theoretical models in which households substitute away from purchases of market insurance into self-insurance via wealth accumulation, it is also consistent with a household’s self-insurance capacity emerging as a by-product of wealth accumulation (for other purposes) over the life cycle. This result is in contrast to many empirical studies that find that life insurance demand increases with the level of household wealth. Our results suggest that in contrast to previous historical studies, life insurance was a sought asset because of its role in the household portfolio during the course of the life cycle. Our findings also indicate that studies examining the impact of government insurance programs on private market insurance demand and savings should account for a household’s accumulated wealth. As household wealth rises, both market insurance demand and precautionary savings decline. If government programs are initiated in labor markets where households have higher levels of wealth, then the government program could give the appearance of crowding out private initiative.

Our study of life insurance demand in 1892 demonstrates for all but the wealthiest of men, there was a negative effect of wealth on the demand for life insurance. Consistent with the theoretical literature on the demand for life insurance, self-insurance, that becomes available to households as a by-product of wealth accumulation, was a substitute for market purchases of life insurance. Our evidence suggests that households primarily demanded life insurance when they lacked accumulated reserves, or wealth, for self-insurance. These conditions are highly correlated with the early stages of the economic life cycle. Our findings suggest that, historically, the indifference many consumers showed towards life insurance has little to do with life insurance being a suspect, immoral or unsought asset. Instead, life insurance was an asset demanded by households at particular points in their life cycles, or by households that were dependent on one earner’s human capital. Finally, there are the broader economic implications for 19th century economic and financial development. During the 19th century, Canada went from being a relatively young frontier society to an older, more settled economy. In 1871, for example, 53% of the population was under age 20 whereas by 1911, this proportion had shrunk to 43%. Over the same period, the proportion aged from 20 to 49 grew from 36 to 44% . The growth in 19th century financial intermediaries and life insurance in particular was no doubt influenced by market demand as a large portion of the population moved into the peak years for life insurance demand. Given that farmers and non-urban workers had less reason to demand insurance, the growth of the insurance industry was propelled in part by the increasing importance of cities, industrial and urban work and the rise of single income households.