The Effects of Social Security on Life Insurance Demand by Married Couples

Abstract

The paper develops a simple one period, utility maximizing model of a married couple choosing the amount of life insurance it wants on each of its earners. The model is specialized to predeceasing husbands and tested using tax and probate data on a sample of Wisconsin households. Husband's future earnings are found to increase the demand for insurance on the husband's life. Social Security survivor benefits decrease this demand, while Social Security benefits conditional on the husband's survival, such as retirement benefits, increase it. These two effects largely offset one another. Wives' future earnings are found to increase the demand for life insurance on the husband, contrary to the model's prediction. While life insurance planning often includes discussion of Old Age, Survivors, and Disability Insurance, more popularly known as Social Security, existing life insurance models do not adequately consider the effects of both Social Security retirement and survivor benefits on life insurance demand.1 Survivor benefits diminish the demand for life while retirement benefits increase it. This paper develops and empirically investigates this hypothesis in a simple model that distinguishes these two effects. The overall effect of Social Security on life insurance bears on the more frequently asked question about the effect of Social Security on bequests, since life insurance is a key tool in bequest planning.2 The paper models a household with two potential earners, and also investigates the role of a spouse's earnings in determining insurance demand. Members of a household can share risks, so that the potential future earnings John Fitzgerald is an Assistant Professor in the Bowdoin College Department of Economics. He earned a Ph.D. at The University of Wisconsin-Madison. 'For example [3] and [20] consider income programming models for life insurance demand where only Social Security survivor benefits are considered. Campbell [5] also neglects retirement benefits. 2The effect of Social Security on bequests remains an unsettled question. In a seminal paper, Barro [2] argued that an unfunded, or pay as you go, Social Security system should have no effects on aggregate savings, since parents will adjust their bequests to offset any burden the system puts on their children. David and Menchik [8] empirically test this hypothesis using data on actual bequests, and find no significant effect of net Social Security wealth on bequests. For a review of the vast literature on the effect of Social Security on savings, see [6]. This content downloaded from 157.55.39.144 on Wed, 07 Sep 2016 04:37:31 UTC All use subject to http://about.jstor.org/terms The Effects of Social Security on Life Insurance 87 of one spouse can reduce the need for life insurance on the other. Given the rise in the number of two earner couples in the United States, the role of the spouse's earnings has become increasingly important. This paper develops a simple one period model of a household (a married couple) choosing the amount of life insurance it desires on each of its two potential earners. The approach is similar to Campbell [5] who solved a single earner problem. An explicit utility maximization model is used where either earner may die during the planning period. The model assumes that two Social Security wealth measures are relevant: the Social Security wealth available to the family conditional on an earner, say the husband, surviving the planning period, and the Social Security wealth available to the surviving family conditional on the husband dying during the planning period. These two meassures have offsetting effects on life insurance demand: the wealth conditional on the husband's survival increases the demand for life much like the husband's future earnings, while wealth conditioned on the husband's death (survivor benefits) reduces the demand for life insurance. Since these effects move in opposite directions, the total (unconditional) effect of Social Security wealth on life insurance demand is indeterminate. This paper measures the degree of offset empirically. The remainder of this section briefly reviews past empirical work on life insurance demand. The next section describes a model of household behavior. The rest of the paper empirically investigates the implications of the model. Section II describes the empirical method, including the data used and calculation of the components of household wealth. Section III presents empirical results, and a brief conclusion follows. Several past studies have looked at the total household expenditure on life insurance premiums, and found that household and wealth have positive effects on premium expenditures.3 Interestingly, Duker [9] found that working wife households have lower total premium expenditures than households in which the wife does no market work. While he terms this underconsumption of insurance, this paper's model shows this to be rational sharing of risk within the household. To this author's knowledge, only one other study has empirically investigated the effect of Social Security on life insurance demand.4 In a framework similar to that used below, Campbell [4] finds that a person's aggregated future earnings have a moderate positive effect on the total insurance held on that person, while aggregated Social Security survivor benefits have a moderate negative effect.5 He ignores Social 'Most of these studies use data from various years of the Survey of Consumer Finances from the Survey Research Center at Michigan. They include [18] [17] [15] and [9]. 4Two other studies, [1] and [11], look at life insurance purchases over a specified period of time for households, and the type of insurance purchased. Neither study considers Social Security. 'Campbell uses data gathered during 1963 and 1964 from the Survey of Financial Characteristics of Consumers and the Survey of Changes in Family Finances. He imputes Social Security benefits based on earnings, age, and family size. This content downloaded from 157.55.39.144 on Wed, 07 Sep 2016 04:37:31 UTC All use subject to http://about.jstor.org/terms 88 The Journal of Risk and Insurance Security retirement benefits, and studies only single earner households, thus ignoring the potential for risk sharing.