Dissolving the relationship between divorce laws and divorce rates

International Review of Law and Economics

Volumn 18, Issue 3, 1998, Pages 341-359

  Ellman, Ira Mark ^a   Lohr, Sharon L ^a  

^a ARIZONA STATE UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 0000229357     PISSN: 01448188     EISSN: None     Source Type: Journal    
DOI: 10.1016/S0144-8188(98)00014-3     Document Type: Article

References (42)

1. Willcox, Walter F. (1891). The Divorce Problem: A Study in Statistics, New York: AMS Press.
2. Wright, Carroll D. (1889). A Report on Marriage and Divorce in the United States, 1867 to 1886. Washington, D.C.: U.S. Government Printing Office (1889).
3. Wright, p. 150.
4. Willcox, p. 55.
5. Friedman, L.M. (1985). A History of American Law (2nd ed.). 204-207, 498-504, New York: Touchstone Books.
6. Plateris, Alexander A. (1973). 100 Years of Marriage and Divorce Statistics, United States, 1867-1967, at 22, Department of HEW Publication number (HRA) 74-1902. Washington, D.C.: U.S. Government Printing Office.
7. See Ellman, Ira Mark, and Sharon Lohr. (1997). "Marriage as Contract, Opportunistic Violence, and Other Bad Arguments for Fault Divorce." University of Illinois Law Review 3:719; and Ellman, Ira Mark. (1997). "The Misguided Movement to Revive Fault Divorce, And Why Reformers Should Look Instead to The American Law Institute." International Journal of Law, Policy and Family 11:216.
8. See Ellman, Ira Mark, and Sharon Lohr. (1997). "Marriage as Contract, Opportunistic Violence, and Other Bad Arguments for Fault Divorce." University of Illinois Law Review 3:719; and Ellman, Ira Mark. (1997). "The Misguided Movement to Revive Fault Divorce, And Why Reformers Should Look Instead to The American Law Institute." International Journal of Law, Policy and Family 11:216.
9. The authorities are systematically described in Ellman, Ira Mark. (1996). "The Place of Fault in a Modern Divorce Law." Arizona State Law Journal 28:773.
10. This issue, pages 325-340.
11. Actually, Brinig and Buckley's thesis is not entirely clear. Although they say that "one would expect less fault and fewer divorces when fault bears a financial penalty," they are not clear as to why one would have that expectation. They do suggest at one point that fault rules impose a financial penalty that will deter marital misbehavior, and one might surmise that they also believe that the spouses are therefore less likely to seek divorce. If that surmise is correct, it would seem that their thesis is multistep: (1) fault rules contain financial penalties for fault, (2) which deter marital misconduct, (3) which in turn reduces the incidence of divorce. Although this seems to be the most plausible reading of their thesis (we cannot discern any other reading that makes any sense at all), it is also a reading that is inconsistent with their classification of states as "fault" and "no-fault." Their statistical analysis, if it is to support their thesis, must classify as a "fault" state only those states having divorce laws that impose a financial penalty on marital misconduct. Yet their analysis treats five states - Arkansas, Utah, Maine, Alaska and New Mexico - as fault states even though their laws bar the consideration of marital misconduct in alimony and property allocation. Their explanation is that because these states permit a spouse to seek divorce on fault as well as no-fault grounds, they "permit the innocent party to introduce fault into the settlement negotiations." But anyone can always introduce anything into the settlement negotiations: the question is whether the availability of fault divorce grounds gives the introduced issue some bite that it would not otherwise have. It is unclear how claims of misconduct could have any bite in these five states.
12. In considering the power of the incentives, one might believe that income effects are likely. That is, one might assume that the same dollar change in alimony would affect the motivations of a poorer spouse more than a richer one. Men are usually better off than their wives. If one also believed that they are more often guilty of legally recognized fault than are their wives, then the resulting imbalance in the incentives would yield more, not fewer, divorces.
13. See Brinig, Margaret F., and Steven M. Crafton. (1994). "Marriage and Opportunism." Journal of Legal Studies 23:869.
14. See Ellman and Lohr, supra note 7.
15. See Ellman, supra note 8.
16. See the studies cited in American Law Institute, Principles of the Law of Family Dissolution (proposed final draft, part I, 1997) at 302.
17. We rely, in the following description of existing fault law, on Ellman, supra note 8.
18. Many of these are cited in Peters, H. Elizabeth (1986). "Marriage and Divorce: Informational Constraints and Private Contracting." American Economic Review 76:437, 448. For a more recent compilation and summary, see Ellman, supra note 7, at 216, 220, 238, note 23.
19. Many of these are cited in Peters, H. Elizabeth (1986). "Marriage and Divorce: Informational Constraints and Private Contracting." American Economic Review 76:437, 448. For a more recent compilation and summary, see Ellman, supra note 7, at 216, 220, 238, note 23.
20. Wright, Gerald C., and Dorothy M. Stetson. (1978). "The Impact of No-Fault Divorce Law Reform on Divorce in American States." Journal of Marriage and Family 40:575, at 580. The exceptions that they found to this rule were California and Florida.
21. Nakonezny, Paul A., Robert D. Shull, and Joseph Lee Rodgers. (1995). "The Effect of No-Fault Divorce Law on the Divorce Rate Across the 50 States and Its Relation to Income, Education, and Religiosity." Journal of Marriage and the Family 57:477.
22. Nakonezny et al., supra note 19 at 485.
23. There are also has a number of other errors; for example, the dates of changes in the law are incorrect for a number of states. We discuss another shortcoming of their analysis in Section IV. Another criticism of the causal argument of Nakonezny et al. is given in Glenn, Norval D. A Reconsideration of the Effect of No-Fault Divorce on Divorce Rates." Journal of Marriage and the Family 59:1023. Glenn shows that for states adopting no-fault divorce outside of the "divorce boom" period of the 1970s, "the mean divorce rate was no higher in the 3 years after adoption than in the 3 years before adoption" (p. 1023).
24. p-Value < 0.0001 for paired t-test. The same holds true for comparing states that did not change their divorce laws in any successive 3-year periods that include 1971. Thus, the average divorce rate is higher for 1970 to 1972 than for 1967 to 1969, and is higher for 1974 to 1976 than for 1971 to 1973.
25. Nakonezny et al. do not identify the particular 6-year period chosen for the comparison done for each state. We surmise that any period between 1950 and 1990 (the years for which they analyze the data) might have been chosen, so long as it did not overlap that particular state's change in divorce law. Arizona changed its divorce law in 1973, during the societal changes of the early 1970s. So, their analysis might have relied upon a 6-year period surrounding 1954, or 1963, or 1985. Yet these as well as many other possible 6-year periods cannot serve as a control for estimating the effect of Arizona's divorce law, because other factors that might have influenced the divorce rate and that might have prompted the change in law were not present during those periods.
26. Kmenta, Jan. (1986). Elements of Econometrics (2nd ed.). New York: Macmillan, p. 616.
27. i in the model would adjust for the mean divorce rate of each state, and hence also adjust for regional effects. If one wanted to account for population mobility, we believe that a much better variable to examine would be immigration to the state. The Spearman correlation coefficient between ENTRY and the average divorce rate from 1980 to 1985 is 0.48; the correlation between (number of immigrants to state between 1970 and 1980)/(1970 population of state) and the average divorce rate from 1980 to 1985 is 0.76 (omitting Louisiana and Nevada). The latter correlation is remarkably high for a social science statistic. We plan to consider its importance further in a forthcoming article.
28. The sources of all data were Vital Statistics of the United States (various years), published by the National Center for Health Statistics, and volumes of Statistical Abstract of the United States, Washington, D.C.: U.S. Government Printing Office.
29. If the population of nonreporting areas is 10% or more of the state population in a given year, as occurred in New Mexico in 1987, Vital Statistics in the United States does not publish the divorce rate. Thus, New Mexico and Indiana both lack data for several years. There is a question of whether the statistical analyses should include divorce rate data from New Mexico and Indiana, because their data are fragmentary. We included data from these states in all analyses reported in Section IV. However, we also analyzed the data without using New Mexico and Indiana, and we found that the omission had no important effect.
30. We note two exceptions: Sepler, Harvey J. (1981). "Measuring the Effects of No-Fault Divorce Laws Across Fifty States: Quantifying a Zeitgeist." Family Law Quarterly 15:65; and Wright, Gerald C., Dorothy M. Stetson. (1978). "The Impact of No-Fault Divorce Law Reform on Divorce in American States." Journal of Marriage and the Family 40:575-580. Sepler plotted divorce rates through 1979 for states using a logarithmic scale. Wright and Stetson plotted adjusted divorce rates for the states that changed their divorce law in 1971, and for California and Florida separately.
31. We note two exceptions: Sepler, Harvey J. (1981). "Measuring the Effects of No-Fault Divorce Laws Across Fifty States: Quantifying a Zeitgeist." Family Law Quarterly 15:65; and Wright, Gerald C., Dorothy M. Stetson. (1978). "The Impact of No-Fault Divorce Law Reform on Divorce in American States." Journal of Marriage and the Family 40:575-580. Sepler plotted divorce rates through 1979 for states using a logarithmic scale. Wright and Stetson plotted adjusted divorce rates for the states that changed their divorce law in 1971, and for California and Florida separately.
32. In "Statistics of Divorce." (1909). Journal of the American Statistical Association 11:486, 491, Joseph A. Hill noted: "The divorce rate is increasing in all parts of the United States; but in the East the increase is comparatively slow, while in the West it is much more rapid, so that the two sections appear to be drawing farther apart in this respect instead of nearer together."
33. Keller, Benjamin. (1889). "Divorces in France." Journal of the American Statistical Association 1:469: "As a natural result of these laws one would expect to find an abnormal number of absolute divorces granted in the first years of their existence, for the statistics for those years would naturally include not only "divorces directs," as the French call those which have not been preceded by a limited divorce or separation de corps, but also a large proportion of decrees granted to persons who had previously obtained limited divorces, and who, unless reconciled, would obviously desire to have such separations converted into absolute divorces," (p. 469).
34. Jane Gentleman and Evelyn Park. (1996). "Patterns of Divorce in Canada: 1970-1993." Proceedings of the Survey Methods Section of the Statistical Society of Canada, Ottawa, Canada, at p. 197-202. The statistics quoted here are on p. 199.
35. Box, George E.P., and George, Tiao. (1975). "Intervention Analysis with Applications to Economic and Environmental Problems." Journal of the American Statistical Association 70:70-79.
36. The west includes states in the Pacific and Mountain regions in Figure 1; north central includes west north central and east north central; south includes west south central, east south central, and south atlantic; northeast includes New England and Middle Atlantic.
37. We started with a default weight of 1. For the analysis of California, for example, we started by giving all years for Arizona weight 1. But Arizona changed its law in 1973: We did not want to have the short-term impact of Arizona's law, if such existed, dominating the overall trend line; we also did not want to exclude Arizona entirely from the trend estimation. As a compromise, we assigned the divorce rate in Arizona weight 0 for 1973; weight 0.2 for 1972 and 1974; weight 0.4 for 1971 and 1975; weight 0.6 for 1970 and 1976; and weight 0.8 for 1969 and 1977. By having a gradually increasing weight function, we avoid jumps in the smoothed trend that would result from suddenly reintroducing a state with a high average divorce rate into the data. Weights were assigned to the data from the other states in the west in a similar fashion to Arizona. In this way, we were able to estimate the trend in the overall divorce rates in the west, without depending on data from California and without relying heavily on data in years surrounding a change in the divorce laws. This was just one possible weighting that could be used; results of the intervention analyses were almost identical with different weighting schemes. In the northeast, for example, it was possible to estimate the trend line with just states that did not change their laws at all during the time period. In the west, however, every state changed its divorce law between 1969 and 1990; if we completely excluded states that changed their laws from the trend line, we would not be able to do an analysis.
38. Friedman, Jerome H. (1984). "A Variable Span Smoother." Technical Report No. 5, Laboratory for Computational Statistics, Dept. of Statistics, Stanford University, California. The Super Smoother uses local cross-validation to choose the bandwidth of the nonparametric smoother, and it is implemented in S-PLUS. See S-PLUS 4 Guide to Statistics, Data Analysis Products Division, MathSoft, Seattle, 1997, p. 160.
39. Friedman, Jerome H. (1984). "A Variable Span Smoother." Technical Report No. 5, Laboratory for Computational Statistics, Dept. of Statistics, Stanford University, California. The Super Smoother uses local cross-validation to choose the bandwidth of the nonparametric smoother, and it is implemented in S-PLUS. See S-PLUS 4 Guide to Statistics, Data Analysis Products Division, MathSoft, Seattle, 1997, p. 160.
40. A discussion of possible factors and a literature review is given in Cameron, Sam. (1995). "A Review of Economic Research into Determinants of Divorce." British Review of Economic Issues 17:1-22.
41. 2 0.41 ± 0.16.
42. Goode, William J. (1993). World Changes in Divorce Patterns, New Haven: Yale University Press, p. 322.