Determinants of youth suicide: the Easterlin-Holinger cohort hypothesis re-examined

American Journal of Economics and Sociology, The,  April, 1998  by Donald G. Freeman

• E-mail
• Print
• Link

All reported equations have adjusted R-squares in excess of .98, indicating a reasonably good fit of the regression line to the data. After the AR1 correction, Durbin-Watson statistics were sufficient to reject the hypothesis of further serial correlation in the residuals.

Interpretation of the Coefficients

Three regressors were consistently robust: relative cohort size, per capita income and percent of families headed by single parents. Coefficient estimates of these regressors remain relatively stable and significant at the five-percent level in combination with any other regressor.

Per capita income has the largest coefficient. A one-percent change in the growth of income, all other factors constant, is associated with about a one-percent increase in the youth suicide rate. This is both larger than and a reversal of sign from Yang's estimate but is consistent with the findings of Hamermesh and Soss. Without further research into the changes in adolescent attitudes and behaviors associated with greater incomes, it is difficult (especially for an economist) to explain why greater affluence leads to feelings of hopelessness. Some possibilities include: 1) inequity in the distribution of this greater income engenders frustration and greater relative despair; 2) ever higher incomes "raise the bar" of success, causing adolescents to lose hope of achieving a living standard commensurate with that of their parents; and 3) greater affluence from parental work effort comes at the expense of less time spent in child nurturing and development.

Changing family structure is another important determinant of the increase in youth suicide. The elasticity of suicide rates with respect to single-parent households is between four-tenths and one-half. The size of this coefficient is fairly large, and its robustness in the presence of changes in poverty suggests that more than relative income differences are at work here. This variable may also be serving as a proxy for higher rates of divorce, teenage pregnancy, or diminished parental influence or control. Single parents will have less time available for nurturing, and this may contribute to a young person's feelings of isolation or hopelessness.

The coefficient for relative cohort size is positive and statistically significant. A one-percent increase in the relative proportion of adolescents to the total population results in a three-tenths to one-third-percent increase in the suicide rate. Arguably the most important finding in this paper, this result demonstrates that controlling for income, family structure and change in economic status reconfirms the Easterlin-Holinger hypothesis of higher risk of youth suicide associated with larger relative cohort size. The disturbing aspect of this finding is that the baby boom "echo" generation is moving into its adolescent years, ending the twenty-year decline in relative cohort size. Even if other factors contributing to increases in suicidal behavior were to stabilize, expanding cohort size may cause a significant increase in youth suicide rates, as will be shown in the following section.