Printer Friendly

The demand for life insurance in Mexico and the United States: a comparative study.

ABSTRACT

This study compares the demand for life insurance in Mexico with that in the United States. It provides a brief historical perspective on the growth of life insurance purchases in the two countries and employs regression analysis to estimate life insurance demand functions. The principal findings are that age, education, and level of income affect the demand for life insurance and that the income elasticity of demand for life insurance is much higher in Mexico than in the United States.

Introduction

The primary purpose of this article is to investigate the factors determining the quantity of life insurance demanded on an aggregate basis using time series data on both Mexico and the United States and to compare the results for the two countries. Most other empirical research on the demand for life insurance in the United States has concentrated on certain population subgroups and utilized cross sectional data. [1] In addition, there has been only limited investigation of the demand for life insurance in less developed countries. As this analysis will show, one apparent difference between the demand for life insurance in the United States and Mexico is that the income elasticity of demand is much higher in Mexico.

A second feature of this article is a brief description and comparison of the growth of life insurance purchases during the last two decades in both Mexico and the United States. This task is undertaken in the next section. Next, the nature of the demand function to be estimated is discussed, and the regression results are presented. The findings of the article are then summarized.

A Brief Historical Perspective

The demand for life insurance in Mexico grew quite rapidly relative to the demand in the United States over the 20-year period from 1964 through 1984. For example, the number of individual life insurance policies in force in Mexico grew from 273,705 in 1964 to 1,561,428 in 1984, a compound annual growth rate of 9.1 percent. The number of group life insurance policies grew even more rapidly, at a compound annual rate of 11.I percent. However, the number of group certificates (individuals insured through group policies) increased at only an 8.8 percent annual rate. In contrast, the number of life insurance policies in the United States grew at only a 1.1 percent compound annual rate over this same period. [2]

The total amount insured (face value of the policies) grew far more rapidly than did the number of policies in Mexico. This figure grew at a compound annual rate of 28.4 percent for individual policies and 32.3 percent for group policies. Part of this exceptional growth was apparently a reaction to rapidly accelerating prices in Mexico, particularly during the 1980s. However, the growth in the Mexican consumer price index combined with that in the economically active population in Mexico certainly cannot account for a major portion of the growth in the amount insured from 1964 to 1980. During this shorter period of relatively stable prices, the amount insured by individual life insurance policies still grew at a 22.4 percent compound annual rate. The corresponding figure for group policies was 28.3 percent. During this same period, the Mexican CPI grew at only an 11.5 percent compound annual rate and the economically active population at a 3.2 percent rate. If these two factors were combined, they could at most account for a 15 percent compound annual rate of growth.

Again, the data for the United States provide a striking contrast to that for Mexico. The average size of an individual life insurance policy in the United States grew at a compound annual rate of only 6.4 percent over the period from 1964 to 1980. The corresponding figure for group policies was 6.9 percent. Moreover, the consumer price index grew at a compound annual rate of 6.3 percent over the same period, so the average size of an individual life insurance policy in the U.S. barely kept up with inflation.

The Regression Model

This section addresses the primary analytical objective of estimating demand functions for life insurance for Mexico and the United States and to compare the results for the two countries. The nature of the estimated demand functions is explained immediately below.

On an abstract level the demand for life insurance, like that for other goods and services, is generally assumed to be a function of its price, a consumer's level of income, the prices and availability of substitute goods, and other personal and environmental characteristics that make life insurance more valuable to the purchaser. A great deal has been written regarding the theoretical basis of the demand for insurance in general and for life insurance in particular. The optimal amount of life insurance is specifically argued to be a function of the household's risk aversion, the accumulated marketable wealth of the household, the household's "intensity for bequests," and the perceived "loading charge," the fraction of the total premium that is in excess of the true expected loss. [3] The household's intensity for bequests is hypothesized to be a function of the age and number of dependents in the household, their future need for economic support, the probability of their future deaths, and the psychological traits of the family, including a sense of moral responsibility and education (See, for example, Campbell, 1980, p. 1162).

On a microeconomic scale, researchers have had some success measuring risk aversion on an individual level. However, Greene (1963 and 1964) found that predicting insurance-buying behavior from risk-taking behavior in other areas (such as gambling) is more difficult. Since this study deals with the demand for insurance on an aggregate level, a relevant measure of "insurance-mindedness" is not available.

As mentioned above, the price of a good or service is usually thought to be a factor affecting the quantity demanded of that good or service. Although several attempts have been made to develop a price index for life insurance, no such data were generally available for the entire period covered by the present study. [4] Therefore, this variable had to be omitted from the estimated demand functions. Aggregate wealth data on a per household basis also were not available over the entire period covered by the present study for either Mexico or the United States so this variable was omitted as well. However, real per capita GNP (gross domestic product or GDP for Mexico) was utilized as a substitute variable in the estimated demand functions for both countries.

As a result, the estimated demand function for insurance was of the following form:

[Mathematical Expression Omitted]

The quantity demanded of life insurance was represented by the average dollar amount (in thousands of dollars) of individual life insurance per family for the United States and the total peso amount (in thousands of pesos) of individual life insurance in force divided by the economically active population of Mexico. Two different variables were used to depict the age distribution of the population for the United States: the median age of the population and the percent of the population between 25 and 64 years of age. Annual data regarding the age distribution of the population for Mexico were not available. Median school years completed was used to reflect the level of education in the United States. Similar education data were not available for Mexico, so a proxy variable equal to the sum of the degrees granted during the year by the Universidad Nacional Autonoma de Mexico and the Instituto Politecnico Nacional was employed.

It was hypothesized that over the relevant ranges of the aggregate data utilized in this study that [Beta]1 [Beta]2, and [Beta]3 would be positive. One could argue that after some age, the quantity demanded of life insurance would be inversely related to the age of the insured. However, this point probably would occur relatively late in life, and even then many people would still have life insurance in force through paid-up policies (See Miller, 1985, p. 34). Moreover, in this study the median age of the U.S. population never was above 32 years of age, and the second age variable utilized, percent of the population between ages 25 and 64, was designed to capture that segment of the population according to age whose demand for life insurance would be the greatest. People in this age bracket probably have the greatest need to protect spouses and/or dependent children from declining incomes as a result of the death of a primary wage earner in the family.

It was further hypothesized that the higher the level of education of the general population, ceteris paribus, the greater would be the demand for life insurance. More highly educated people would recognize the various types of life insurance available and perhaps have a stronger desire to protect dependents in this way. This latter inclination is related to the "intensity for bequest" motive discussed above.

Finally, for the relevant income levels included in this study, it was hypothesized that on an aggregate basis the demand for insurance would be positively related to income. Most people in either the United States or Mexico are not so wealthy that their dependents would perceive no substantial decline in income if the primary family provider(s) were to die. Thus, life insurance is generally considered to be a desirable good to protect dependents, and the higher the income of the family member (or members) who supports the family, the greater the demand for life insurance to protect this standard of living.

The demand functions were estimated using a current value of real per capita GNP and a three-period future value for this variable. The future value of the income variable was also used because it was hypothesized that it would be reasonable that those supporting a family would consider their anticipated level of income and in their desire to protect their dependents from loss of support. [5] Annual data covering the period from 1960 to 1982 for the United States and from 1964 to 1979 for Mexico were employed.

Regression Results

The results of the regression analysis are given in Table 1. As can be seen from that table, all of the estimated coefficients of the independent variables had the hypothesized sign, and all but two were significantly greater than zero at the 2.5 percent level of significance in the relationships for both countries. The coefficient of the education variable was significantly greater than zero at only the 10 percent level of significance in the estimated demand function for Mexico with future per capita GDP. Also, the estimated coefficient of the real per capita GNP variable using the second age variable for the United States was significantly greater than zero at only the 5 percent level of significance. The R[.sub.2] values ranged from .97 for the estimated functions for Mexico to .98 and .99 for the United States. The Durbin-Watson statistic was consistent with the hypothesis that no serial correlation was present at the 5 percent level of significance in the estimated demand function for Mexico using future per capita GNP as an independent variable. It was in the inconclusive range for the other estimated demand function for Mexico and the two versions of the estimated demand function using future per capita GNP for the United States. The Durbin-Watson statistic for the two estimated demand functions for the United States that used current per capita GNP as an independent variable did indicate the presence of serial correlation at the 5 percent level of significance. These results would be consistent with the hypothesis that the demand function with a future value of per capita GNP as an independent variable was a more appropriate model than that with the current value of per capita GNP.

Perhaps the most striking observation one can make when comparing the estimated demand functions for Mexico and the United States is that the estimated income elasticity of demand for life insurance is much greater for Mexico than for the United States. In fact, this estimated elasticity was over three times as great for Mexico as for the United States in all corresponding cases. This result would be consistent with the hypothesis that the income elasticity of demand for life insurance is much higher at lower levels of income than at higher income levels. Such a situation would seem reasonable since a high income family would likely have accumulated greater wealth to protect the standard of living of the family if a major income-earner died.

The estimates of the income elasticity of demand for life insurance in the United States are similar to those obtained by Hammond, Houston, and Melander (1967) using cross sectional data for 1962. Except for a low income group, their estimates of income elasticity of demand ranged from 0.61 to 1.53. They also found that while the income elasticity of demand was greater for a middle income group than for a lower income group, it was also much lower for a higher income group than for the middle income group. Babbel (1985, p. 233) obtained estimates of the income elasticity of demand for life insurance that ranged between .68 and 1.09.

Hammond, Houston, and Melander (1967) also found education to be significant at the 5 percent level in some of their equations and not significant in others. This finding may have been the result of collinearity between their education and occupation variables. The latter variable was also included as an independent variable, and it was significant in all of their estimated relationships at the 5 percent level of significance. They found age to be a significant factor affecting premium expenditures for life insurance for the low income and middle income groups. However, the estimated coefficient of this variable had a negative sign for the low income group. [6]

In another cross sectional study, Duker (1967, p. 528) found that income, occupation, education, total assets, and age had "significant partial regression coefficients." However, since Duker was concerned primarily with the effect of a spouse's employment on family expenditures for life insurance premiums, he did not report these results in detail. Finally, in a more recent study that included both psychographic as well as demographic variables, Burnett and Palmer (1984) found education, number of children, and income to be the best demographic predictors of the demand for life insurance.

Conclusions

The demand estimation results in this article are generally consistent with the hypothesis that age, education, and level of income are factors that affect the demand for life insurance. These demand function estimates are also consistent with the hypothesis that the income elasticity of demand for life insurance in Mexico is far higher than that for the United States. This finding is consistent with those of other researchers which have indicated a higher income elasticity of demand for life insurance at lower income levels than at much higher income levels.

1 See Anderson and Nevin, 1975; Babbel, 1981; Berekson, 1972; Briys and Louberge, 1985; and Duker, 1967.

2 The data sources for this study include Nacional Financiera, s.a., La economia mexicana en cifras, Mexico, D.F., 1981; Secretaria de Programacion y Presupuesto, Estadisticas historicas de Mexico, Mexico, D.F.: Instituto Nacional de Estadistica, Geografia e Informitica, 1985; Secretaria de Programacion y Presupuesto, Annuario estadistico de los Estados Unidos Mexicanos, Mexico, D.F.: Instituto Nacional de Estadistica, Geografia e Informitica, various issues; U.S. Bureau of the Census, Current Population Reports, Series P-20; U.S. Bureau of the Census, Statistical Abstract of the United States, Washington, D.C.: U.S. Government Printing Office, various issues; U.S. Bureau of the Census, Historical Statistics of the United States: Colonial Times to 1970, Washington, D.C.: U.S. Government Printing Office, 1975; and Economic Report of the President, Washington, D.C.: U.S. Government Printing Office, 1987.

3 See Campbell, 1980. Also see Abel, 1986; Arrow, 1974; Briys and Louberge, 1985; Karni and Zilcha, 1986; Kotlikoff, 1986; Mossin, 1968; and Smith, 1968.

4 See Babbel, 1985; Babbel and Staking, 1983; and Lin, 1971.

5 Babbel, 1985, used a three-year moving average as a proxy for "permanent" income in his study.

6 Also see the results of a study by Berekson, 1972. Since this latter study involved students on two college campuses, its results may be less indicative of the demand for insurance on the part of adults in the United States as a group.

REFERENCES

1. Abel, Andrew B., 1986, Comment on Kotlikoff, Shoven, and Spivak,

Journal of Labor Economics, 4: S208-S215.

2. Anderson, Dan R., and John R. Nevin, 1975, Determinants of Young

Marrieds' Life Insurance Purchasing Behavior: An Empirical Investigation,

The Journal of Risk and Insurance, 42: 375-87.

3. Arrow, Kenneth J., 1974, Optimal Insurance Purchasing and Generalized

Deductibles, Scandinavian Actuarial Journal, 1-42.

4. Babbel, David F., 1981, Inflation, Indexation, and Life Insurance Sales in

Brazil, The Journal of Risk and Insurance, 48: 111-35.

5. __ 1985, The Price Elasticity of Demand for Whole Life Insurance,

The Journal of Finance, 40: 225-39.

6. __and K. B. Staking, 1983, A Capital Budgeting Analysis of Life

Insurance Costs in the United States: 1950-1979, The Journal of Finance,

38: 149-70.

7. Berekson, Leonard L., 1972, Birth Order, Anxiety, Affiliation, and the

Purchase of Life Insurance, The Journal of Risk and Insurance, 39:

93-108.

8. Briys, Eric P., and Henri Louberge, 1985, On the Theory of Rational

Insurance Purchasing: A Note, The Journal of Finance, 40: 577-81.

9. Burnett, John J., and Bruce A. Palmer, 1984, Examining Life Insurance

Ownership Through Demographic and Psychographic Characteristics,

The Journal of Risk and Insurance, 51: 453-67.

10. Campbell, Ritchie A., 1980, The Demand for Life Insurance: An

Application of the Economics of Uncertainty, The Journal of Finance, 35:

1155-72.

11. Duker, Jacob M., 1967, Expenditures for Life Insurance Among

Working-Wife Families, The Journal of Risk and Insurance, 36: 525-33.

12. Ferber, Robert, and Lucy Chao Lee, 1980, Acquisition and Accumulation

of Life Insurance in Early Married Life, The Journal of Risk and

Insurance, 47: 713-34.

13. Greene, Mark R., 1963, Attitudes Toward Risk and A Theory of

Insurance Consumption, The Journal of Insurance, 30: 165-82.

14. __ 1964, 'Insurance Mindedness'- Implications for Insurance

Theory, The Journal of Risk and Insurance, 31: 27-38.

15. Hammond, J.D., David B. Houston, and Eugene R. Melander, 1967,

Determinants of Household Life Insurance Premium Expenditures: An

Empirical Investigation, The Journal of Risk and Insurance, 34: 397-408.

16. Karni, Edi, and Itzhak Zilcha, 1986, Risk Aversion in the Theory of Life

Insurance: The Fisherian Model, The Journal o Risk and Insurance, 53:

606-20.

17. Kotlikoff, Laurence J., John Shoven, and Avia Spivak, 1986, The Effect

of Annuity Insurance on Savings and Inequality, Journal of Labor

Economics, 4: S183-S207.

18. Lin, Cheyeh, 1971, A Price Index for Life Insurance, The Journal of Risk

and Insurance, 38: 563-70.

19. Miller, Michael A., 1985, Age-Related Reductions in Worker's Life

Insurance, Monthly Labor Review, September 1985, 29-34.

20. Mossin, Jan, 1968, Aspects of Rational Insurance Purchasing, Journal of

Political Economy, 76: 553-68.

21. Smith, Vernon L., 1968, Optimal Insurance Coverage, Journal of Political

Economy, 76: 68-77.

Dale B. Truett is Professor of Economics and Finance and Lila J. Truett is Professor and Director of the Division of Economics and Finance at the University of Texas at San Antonio.
COPYRIGHT 1990 American Risk and Insurance Association, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1990 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Truett, Dale B.; Turett, Lila J.
Publication:Journal of Risk and Insurance
Date:Jun 1, 1990
Words:3220
Previous Article:Karl Borch's research contributions to insurance.
Next Article:Moral hazard and moral imperative.
Topics:

Terms of use | Copyright © 2017 Farlex, Inc. | Feedback | For webmasters