BARGAINING IN THE SHADOW OF THE LAW: DIVORCE LAWS …users.nber.org/~jwolfers/Papers/bargaining_in_the_shadow_of_the_law.pdfunilateral divorce in California. This legislative change

BARGAINING IN THE SHADOW OF THE LAWDIVORCE LAWS AND FAMILY DISTRESS

BETSEY STEVENSON AND JUSTIN WOLFERS

This paper exploits the variation occurring from the different timing ofdivorce law reforms across the United States to evaluate how unilateral divorcechanged family violence and whether the option provided by unilateral divorcereduced suicide and spousal homicide Unilateral divorce both potentially in-creases the likelihood that a domestic violence relationship ends and acts totransfer bargaining power toward the abused thereby potentially stopping theabuse in extant relationships In states that introduced unilateral divorce we finda 8ndash16 percent decline in female suicide roughly a 30 percent decline in domesticviolence for both men and women and a 10 percent decline in females murderedby their partners

I INTRODUCTION

In 1969 then Governor Ronald Reagan signed a bill creatingunilateral divorce in California This legislative change was oneof the first in a series that increased access to divorce across thenation During the next two decades many states moved awayfrom fault-based divorces which were challenging the legal sys-tem toward the less adversarial unilateral divorce1 In otherwords in many states it became possible to seek the dissolution ofa marriage without the consent of onersquos spouse

However many states began to question these changes in the

This project has drawn on the advice of many generous friends and col-leagues including Olivier Blanchard Margaret Brinig David Cutler ThomasDee David Ellwood Leora Friedberg Edward Glaeser Claudia Goldin CarolineMinter Hoxby Christopher Jencks Alan Krueger Steven Levitt Jeffrey MironKatherine Newman Robert Putnam David Weiman Julie Wilson and seminarparticipants at Harvard University the MacArthur Network on Inequality andSocial Interactions Stanford University University of Michigan Princeton Uni-versity the London School of Economics University of California at BerkeleyColumbia University Yale University University of Melbourne and the Societyof Labor Economists Special thanks goes to Lawrence Katz for his guidancethroughout the project We have also benefited from the excellent research assis-tance of Eric Klotch Amalia Miller and Jason Grissom Remaining errors are ourown We are grateful to the MacArthur Foundation and the Social Science Re-search Council for funding for this project A previous version of this paper wascirculated under the title ldquorsquoTil Death Do Us Part Effects of Divorce Laws onSuicide Domestic Violence and Intimate Homiciderdquo

1 Historical accounts of the legislative movement to pass unilateral divorcefocus on the difficulty of an adversarial system in which fault-based grounds fordivorce need to be proved While legitimate cases sometimes struggled to establishsufficient evidence in the face of a denying spouse cases in which both coupleswanted to divorce often involved fraudulent charges of adultery and abuse asspouses attempted to convince the court (usually successfully) that these werelegitimate grounds for divorce [Jacob 1988]

copy 2006 by the President and Fellows of Harvard College and the Massachusetts Institute ofTechnologyThe Quarterly Journal of Economics February 2006

267

1990s and 2000s Widespread concern over the decline of theAmerican family has led many to point the finger at unilateraldivorce laws claiming that easy access to divorce underminestraditional family structures Unfortunately much of the publicdiscussion centers on the consequences of divorce rather than theconsequences of divorce laws

Unilateral divorce allows marriages to end where one personwants out of the marriage and the other person wants to remainmarried This paper seeks to answer the question who benefitedfrom this change and by how much While models of the familythat rely on a common preference function or internal threatpoints predict little change external threat-point models tell usthat unilateral divorce changes bargaining within marriage byimproving the outside options of each spouse As such bargainingpower and therefore resources shifts toward the person whomost wants out of the marriage

The people most likely to benefit from unilateral divorce aretherefore those who stand to gain the most from having the optionto exit their relationship One possibility is that those in violentpotentially lethal relationships have the most to gain when theycan credibly threaten to exit the relationships Unilateral divorcehas two possible effects on these relationships The first is that itallows them to end2 The second is that the threat of divorce maybe sufficient to prevent future abuse in relationships that con-tinue Our focus in this paper is the effect of allowing unilateraldivorce on these particularly bad marriages potentially throughboth channels

Without access to unilateral divorce people ldquotrappedrdquo in abad marriage had few choices While they could leave the mar-riage without being granted a divorce they would not be able totake any assets from the marriage and would be unable to re-marry We consider the possibility that violent relationships weremore likely to end through suicide or homicide prior to unilateraldivorce Suicides could result from unhappiness the value ofcontinuing to live in the abusive relationship falling below theoption value of staying alive3 Alternatively those in abusiverelationships may have used strategic suicide attempts as a way

2 While fault-based divorce does offer divorce for violent relationships theviolence must be proved in court These cases were quite adversarial and manyabuse victims were likely afraid of the heightened threat during the trial

3 For an economic model of suicide see Hamermesh and Soss [1974]

268 QUARTERLY JOURNAL OF ECONOMICS

to get more resources transferred toward them4 Homicide mayresult either because the victim of abuse fights back with lethalforce or because the abuse itself becomes lethal

This paper exploits the variation occurring from the differenttiming of divorce law reforms across the United States to evaluatehow unilateral divorce changed family violence and whether theoption provided by unilateral divorce reduced suicide and spousalhomicide Family violence surveys conducted in the mid-1970sand again in the mid-1980s provide basic detail about domesticviolence by both men and women Spousal homicide and suiciderates are examined for both men and women

We find that states that passed unilateral divorce laws saw alarge decline in both male- and female-initiated domestic vio-lence Between 1976 and 1985 states that had changed theirdivorce laws to allow unilateral divorce saw their overall andsevere domestic violence rates fall by about one-third The effecton domestic violence is large enough to imply that domesticviolence was reduced not just by ending violent relationships butby reducing the violence in extant relationships as well

Our findings examining potential lethal ends to domesticviolencemdashsuicide and homicidemdashpoint to the benefits of unilat-eral divorce for women We show that women murdered by inti-mates declined by 10 percent following the introduction of uni-lateral divorce However we note that an examination of thedynamic effects of the change by year indicate that there mayhave been a preexisting downward trend in women being killedby intimates in states that adopted unilateral divorce We find nodiscernible effect of unilateral divorce laws on spousal homicidefor men

Suicide rates are examined for all men and women sepa-rately and by age category To capture the full dynamic responseof the suicide rate to the law change we evaluate the effect foreach year following the adoption of unilateral divorce As withspousal homicide our results show no discernible effect of unilat-eral divorce on male suicide Female suicide is shown to fallfollowing the adoption of unilateral divorce Furthermore theresults indicate that female suicide rates continue to fall in uni-lateral divorce states for more than a decade following the legalchange Averaging the effects over the twenty years following

4 For a more complete discussion of strategic suicide see Cutler Glaeserand Norberg [2001]

269BARGAINING IN THE SHADOW OF THE LAW

reform suggests an aggregate decline of 5ndash10 percent with largerlong-run effects We now turn to theory to better elucidate the keyforces mediating these results

II MEDIATING FORCES MARRIAGE DIVORCE AND BARGAINING

WITHIN MARRIAGE

Unilateral divorce laws may change behavior through twoprimary channels First they may lead to a change in divorcerates allowing those to escape who were unable to either provefault or persuade their spouse to grant them a divorce Andsecond these laws redistribute property rights and hence bar-gaining power within the relationship Becker [1993] has arguedthat the Coase theorem is the natural starting point for such ananalysis

In a Coasian analysis unilateral divorce laws simply trans-fer a well-defined property rightmdashthe right to remarrymdashfrom thespouse who wants to remain married to the partner desiring adivorce Efficient bargaining ensures that marriages only dissolveif marriage is jointly suboptimal and this efficient bargain will beobtained irrespective of the initial assignment of property rightsAs such the Coase theorem predicts that there are no ldquoinefficientmarriagesrdquo and a change in divorce law to allow unilateral di-vorce will have no effect on the divorce rate Therefore the firsteffect of unilateral divorcemdashallowing certain marriages to endthat would not otherwise have endedmdashonly occurs in cases wherethe Coase theorem is violated5

Research has shown that the divorce rate was affected by thepassage of unilateral divorce Wolfers [2006] finds a small andtransitory rise in divorce that dissipated within a decade How-ever the magnitude of this effect suggests only a very small andgradual change in the stock of married couples6 Yet a smallincrease in divorce could reflect a large proportion of those inviolent relationships divorcing including those that might other-

5 The Coase Theorem requires costless bargaining transferable utility andno wealth effects

6 Combining the estimates in Wolfers [2006] and Rasul [2004] the propor-tion of the population who are married declines by about 1ndash2 percent in the decadefollowing reform (relative to the control states) with the effects becoming onlyslightly larger over the ensuing decade


wise have ended lethally through suicide or homicide A Coasianprediction of no change in the divorce rate requires costless bar-gaining something that seems particularly unlikely to apply tothose marriages where violence (rather than negotiation) is usedto settle conflicting claims By ending inefficient (and violent)marriages unilateral divorce both reduces domestic violence andraises the expected value of life for the partner trapped in aninefficient marriage thus reducing suicide

Domestic violence however comes in varying degrees and alarge decline in overall domestic violence cannot simply be ex-plained by increased divorce over 10 percent of couples acknowl-edge using some amount of violence during a spousal conflictThis leads us to consider the second channel through whichunilateral divorce may impact spousal violence the distributionof bargaining power within marriage

While Coase predicts a change in distribution toward thosewho want out of the marriage (this redistribution is the set of sidepayments required to enforce an efficient bargain) the effects ofredistribution depends on the underlying model of intrahouse-hold distribution Existing theory is conflicted about whether aredistribution of resources within a family will affect individualmembersrsquo shares of resources Both the common preference ap-proach to within-family distribution and internal threat point(separate spheres) bargaining models argue that the change inproperty rights within a marriage should have no effect on with-in-household distribution7 The former rules out spousal bargain-ing by positing a joint utility function (perhaps love yields perfectaltruism and hence a common preference) As such the Coasetheorem predictions about outcomes will hold (the common pref-erence model posits that households maximize a joint utilityfunction and as such divorce rates would be invariant to divorcelaw) however distribution will remain unchanged Internalthreat point models argue that distribution is determinedthrough bargaining however the relevant threat points are re-version to a noncooperative equilibrium (such as sleeping on thecouch) within the marriage and are invariant to a change inoutside options Unilateral divorce laws do not affect these threat

7 For information on bargaining models that rely on threat points that areinternal to the marriage see Lundberg and Pollak [1993]


points and hence do not change the distribution of resourceswithin a household

By contrast exit threat bargaining models emphasize eachspousersquos best option outside the marriage as the relevant parame-ters determining the intrahousehold distribution Under a con-sent divorce regime the relevant exit threat is to leave the mar-riage albeit with no opportunity to remarry nor with a legalclaim to a share of the couplersquos joint assets Unilateral divorcelaws provide for a more attractive outside option which likelyaffects the resulting bargain inside the marriage Alternativelyphrased bargaining power and thus resources should be redis-tributed toward those for whom unilateral divorce provides acredit threat to exit the marriage

If the redistribution of property rights caused by unilateraldivorce laws does change within-household bargaining we shouldsee effects arising out of that redistribution If unilateral divorcelaws redistribute bargaining power toward abused spouses pre-sumably abused spouses will use their increased bargainingpower to demand less abuse Furthermore redistribution shouldhave the largest impact on those for whom the marginal utility ofan extra dollar is the highest Such relationships might involvehighly skewed distribution These are also the relationships inwhich one might expect to observe extreme attempts to redistrib-ute resources Cutler Glaeser and Norberg [2001] suggest thatldquostrategicrdquo suicide attempts may be designed to signal unhappi-ness with the current intrahousehold allocation and to threatenthe abuser with a bad outcome if it is not rectified If the threat issuccessful it leads to a redistribution of resources toward thesuicidal spouse Strategic suicide must be (occasionally) credibleto be effective as a threat and as such must result some propor-tion of the time in actual suicides By transferring bargainingpower toward the person who is enduring violence they can usethis increased power to negotiate less violence As such thisincreased power also reduces the marginal value of strategicsuicide attempts (assuming decreasing returns to lowering vio-lence) thereby reducing both attempts and actual suicides(ldquofailedrdquo attempts)

Finally most spousal homicides occur in the context of abu-sive relationships [Campbell 1992] and hence any policy thatreduces domestic violence is likely to reduce the probability ofspousal homicide We now turn to exploring these potential ef-fects empirically


III EMPIRICAL STRATEGY AND DATA

We follow Friedbergrsquos [1998] coding of state divorce regimesand the dates of divorce reforms8 It should be noted that thereare actually degrees of unilateral divorce in that legislationmight allow unilateral divorce conditional upon a separation pe-riod We code states both with and without separation require-ments as unilateral divorce regimes9 Of the 50 states 5 are yetto adopt any form of unilateral divorce Arkansas DelawareMississippi New York and Tennessee Of the 45 states thatcurrently have unilateral divorce regimes 9 had adopted somevariant of unilateral divorce before the no-fault revolution of theearly 1970s Along with the 36 remaining states we include theDistrict of Columbia which adopted unilateral divorce in 1977 Con-sequently we effectively have 37 ldquoexperimentsrdquo of changing divorcelaws The remaining fourteen states are included as controls

We use the natural variation resulting from the differenttiming of the adoption of unilateral divorce laws across states toestimate the effects of these laws on suicide domestic violenceand homicide rates for women and men independently Conse-quently we use state-based panel estimation including bothstate and time fixed effects in all regressions A dummy variableindicating whether the state currently allows unilateral divorce isour variable of interest The dependent variable is the annualsuicide domestic violence or murder rate Where possible wereport our coefficients as elasticities (evaluated at the unweightedcell mean) That is the reported results are interpreted as thepercentage change in the relevant rate stemming from the changeto unilateral divorce10

Data on suicide come from the National Center for HealthStatistics (NCHS)11 The NCHS data are a census of death cer-

8 Results are consistent with alternative coding of the dates of the legalreforms to unilateral divorce

9 Around one-third of states have separation requirements ranging from sixmonths to five years Results are consistent with alternative treatment of sepa-ration requirements

10 Summary statistics are available in Stevenson and Wolfers [2003]11 Suicide data for 1964ndash1967 were hand entered from annual editions of

the NCHS report ldquoVital Statistics Mortality Vol 2rdquo Data for 1968ndash1978 arecalculated from ICPSR Study No 8224 ldquoMortality Detail Files External CauseExtract 1968ndash78rdquo PI National Center for Health Statistics Data from 1979ndash1996 have been downloaded from the Center for Disease Controlrsquos Wonder systemwhich accesses the NCHS ldquoCompressed Mortality Filesrdquo (httpwondercdcgov)Apart from minor revisions to the International Classification of Diseases thesedata are consistently coded


tificates which code the cause of death for all deceased personsThere are broad codes for suicide as well as a more detailedcoding structure that includes data on the method of suicideIndividual data on gender state of residence and age of death arealso collected

Data on domestic violence are from the landmark FamilyViolence Surveys undertaken by sociologists Murray A Strausand Richard J Gelles in 1976 and again in 198512 These data aregathered using household interviews that ask how couples re-solve conflict This type of survey instrument typically yieldshigher estimates of domestic violence than police reports or crimevictimization surveys because the victim need not perceive the actas domestic violence or a crime for it to be recorded13 While stillan imperfect survey instrument Markowitz [2000 p 286] arguesthat this methodology is currently ldquothe best available techniquefor collecting truthful information on domestic violencerdquo

Data on homicide come from the FBI Uniform Crime Reports(UCR)14 The UCR data are derived using a voluntary policeagency-based reporting system The Supplementary HomicideReports of the UCR provide incident-level information on criminalhomicides including data describing the date and location of theincident as well as a range of information on both the offenderand the victim The particular richness of these data is that itcodes the relationship of the victim to the murderer whereknown

Because the FBI data rely on police reporting there are oftenproblems of underreporting or downgrading of crimes Howeverthe nature of homicide means that both of these problems areminimized The FBI counts of total murders each year by statewere checked against the independently gathered NCHS murdercount Generally these two data sources were consistent and

12 The 1976 and 1985 surveys are ICPSR studies 7733 and 9211respectively

13 Crime victimization survey data lack state identifiers and are not avail-able for the relevant time period Police reports suffer from serious problems ofunderreporting and changes in social norms regarding reporting over the relevanttime period

14 Data for 1968ndash1975 are from ICPSR Study No 8676 ldquoTrends in Ameri-can Homicide 1968ndash1978 Victim-Level Supplementary Homicide Reportsrdquo[Riedel and Zahn 1994] Data for 1976ndash1994 are extracted from ICPSR Study No6754 ldquoUniform Crime Reports [United States] Supplementary Homicide Reports1976ndash1994rdquo [Fox 1996] A detailed appendix discussing the consistency of thesedata is available from the authors


hence the rest of our analysis uses the FBI data which includetheir coding of victim-perpetrator relationships

Nonetheless there remains a range of problems when work-ing with these data First the participation of agencies is notcompletely consistent and when an agency fails to report in aparticular month we cannot tell whether this reflects laxity withpaperwork or that there were no murders to report15 Secondthere are various coding breaks arising from the changing defi-nitions of victim-perpetrator relationship causing a minor breakin 1972 and a more important break in 1976 These codingbreaks present a problem for our analysis because conceptuallywe would like to capture any relationship that may be affected bychanges in family law Such relationships include along withspouses domestic and nondomestic romantic partners and otherfamily members (particularly children) However there are dataproblems constructing such a series that is consistent acrosscoding breaks As such we estimate our results for several defi-nitions of intimate homicide

IV SUICIDE RESULTS

By examining the period from 1964 through to 1996 we canboth robustly identify suicide rates before the adoption of unilat-eral divorce laws and trace their evolution over the followingyears Note that the dependent variable is the suicide rate of allpersons not just those who have been married We analyze thisvariable both because of data limitations (the NCHS begin codingmarital status in 1978) and to avoid endogeneity problems posedby the possibility that marriage decisions may respond to divorceregime By analyzing the suicide rate of all persons our coeffi-cient captures the effect of unilateral divorce on suicidalitythrough both channels those who remain married and those whoexit their relationships

15 When there are no data for an entire state for a whole year this couldreflect either that the state was not participating in the reporting program or thatthere were no murders in that state-year We assume nonparticipation when azero murder count would lie outside a three-standard error confidence band forthat state and infer a number by linear interpolation Otherwise we assume azero murder count These adjustments affect 37 of our 2754 state-year-sex obser-vations One outlier to this is Illinois where the Chicago Police Department failedto report any murders in 1984 1985 November 1986ndashMay 1987 July 1987ndashDecember 1987 and July 1990ndashDecember 1990 As it is implausible that therewere no murders during these periods we omit Illinois from our homicidesamples


We employ OLS to estimate

Suicide ratest k

kUnilateralstk

s

sStates t

tYeart Controlsst εst

Unilateralk refers to a series of dummy variables set equal to oneif a state had adopted unilateral divorce k years ago Thuscoefficients are reported as the percentage change in the suiciderate due to the adoption of unilateral divorce laws the statednumber of years ago As such they map out the full dynamicresponse of the suicide rate to the law change

The first and third columns of Table I report baseline resultswithout including demographic and social policy controls for menand women respectively The second and fourth columns add afull set of controls including a proxy for the evolving economicpower of women (the ratio of male-to-female employment rates)business cycle indicators (state income per capita and unemploy-ment) welfare generosity (the maximum AFDC payment for afamily of four and the share of the state population on the welfarerolls) the availability of abortion and the racial and age compo-sition of the state16 While we find that some of these controls aresignificant explanators of the suicide rate their inclusion haslittle effect on our parameter of interestmdashthe estimated effect ofunilateral divorce

Table I shows that there is a large and statistically signifi-cant reduction in the female suicide rate following the change tounilateral divorce Further this effect grows over time with thefull effects of unilateral divorce on female suicide occurring fifteento twenty years following the adoption of unilateral divorce Av-eraging the effects over the twenty years following reform sug-gests an aggregate decline of 8 percentndash10 percent in femalesuicide and a long-run decline that is much larger For malesuicides Table I reveals no discernible effect It should be notedthat the male suicide rate is four times larger than that forwomen thus these results falsify neither moderately large posi-tive nor negative effects on men committing suicide

We test the sensitivity of our results to a number of alterna-

16 Our population data downloaded from wwwcensusgov are not coded bygender the evolution of gender shares in each state is imputed from the MarchCPS files (for the population aged fourteen or over)


TABLE IEFFECTS OF UNILATERAL DIVORCE ON SUICIDE RATES (PERCENT CHANGE)

Column no

Female suicides Male suicides

(1f) (2f) (1m) (2m)

Year of change 16 13 08 14(38) (34) (22) (21)

1ndash2 years later 15 14 12 05(37) (35) (15) (14)









19 years later 187 164 39 43(32) (33) (20) (20)

Mean suicide rate54 suicides permillion women

202 suicides permillion men

Average effect over the 20years following divorcelaw reform

97(23)

83(23)

15(13)

20(13)

F-test of joint significance p 000 p 000 p 036 p 037Control variablesState and year fixed effects

Economic demographic andsocial policy controls

Sample 1964ndash1996 n 1683Dependent variable is the aggregate state suicide rate by year Coefficients are reported as the percent-

age change in the suicide rate due to the adoption of unilateral divorce laws the stated number of years agothis elasticity is calculated using the unweighted cell mean as the base Robust standard errors are inparentheses

Controls include the maximum AFDC rate for a family of four the natural log of state personal incomeper capita the unemployment rate the female-to-male employment rate age composition variables indicat-ing the share of statesrsquo populations aged 14ndash19 and then ten-year cohorts beginning with age 20 up to avariable for 90 and the share of the statersquos population that is Black White and other (Employment statusage and race data are constructed from Uniconrsquos March CPS files and refer to the population aged fourteenyears or greater)


tive specifications17 We examine the sensitivity of our baselineregressions to the time period and sample chosen by omitting inturn individual states or years finding that particular states oryears do not unduly influence our results Robust estimationprocedures including median regression also yield similar re-sults Further while OLS implicitly gives equal weight to each ofour 37 divorce reform experiments we also found similar resultsusing population-weighted least squares and generalized leastsquares

In further robustness testing we tested our results to theinclusion of state-specific time trends finding that their inclusioncauses the standard errors to increase For women the specifica-tion including state-specific time trends yields point estates thatare roughly similar to but slightly smaller than those shown inTable I However the increase in standard errors yields resultsthat are not precisely estimated enough to reject either a null thatthe pattern of coefficients follows that shown in Table I or a nullof no effect For males including state-specific trends is sugges-tive of a decline in male suicide rates following the advent ofunilateral divorce We also experimented with the control groupdropping those states that did not change their divorce laws fromthe estimation We found that estimating off only the variationdue to the different timing of reform was sufficient to identify thenoted large decline in female suicide This specification was alsosuggestive of a decline in male suicide

Timing evidence might speak to a causal interpretation ofthese results We are particularly interested in whether thechange in suicide postdated the change in divorce regime andwhether adjustment to the new regime seems plausible Addition-ally if divorce law is directly affecting suicidality it should pri-marily affect prime-age women rather than teens and the elderlyIn order to examine these issues we added a series of leads to ourpreferred specification coding dummies for whether unilateraldivorce will become law in 1ndash2 years 3ndash4 years and so on withleads beyond ten years coded to the 9ndash10 year group Again wefind no discernible effect on male suicide For female suicides thecoefficients on the dummies indicating the period prior to thedivorce law reform are all close to zero and in no case are they(individually or jointly) statistically distinguishable from zero

17 Several of the specification tests discussed can be found in Stevenson andWolfers [2003]


We also disaggregate our main results by age Figure I re-ports these regressions for eleven different age groups These agegroups comprise unequal shares of the population and so in eachcase coefficients are scaled by their share of the U S populationallowing these figures to be added to yield the aggregate effect(shown in the bottom right panel) For teens the effect is arelatively precisely estimated zero reflecting both the lack ofcorrelation between teen suicide and divorce laws and the rela-tively small number of teen suicides The second row of Figure Ishows that prime-age women account for the bulk of the maineffect with unilateral divorce substantially reducing the suiciderates of women in each of the age groups from 25ndash65 Turning tothe elderly it appears that unilateral divorce laws had littleeffect on suicide decisions although there may be some impacton women aged 65ndash74 (these estimates are sufficiently impre-cise as to be consistent with either no effect or a meaningfuldecline) Overall the observed correlation between the adop-tion of unilateral divorce and the decline in female suicideseems robust and we can be confident that neither youth northe elderly drive the observed correlation between female sui-cide and divorce regime

V DOMESTIC VIOLENCE AND HOMICIDE

While the Strauss and Gelles [1994] data on domestic vio-lence are plausibly the best available data the timing of thesurveys is not ideal for evaluating the effect of unilateral divorceon domestic violence These surveys provide cross-sectional datafor 1976 by which time 31 states had recently changed theirdivorce laws and again for 1985 by which time 37 states (includ-ing Washington DC) had changed their laws to allow unilateraldivorce This timing is somewhat unfortunate in that it is unclearhow the differential timing of reform across states would trans-late into differential changes in domestic violence rates over the1976ndash1985 period Although the differential cross-state timing inreform yields little analytical leverage we can compare changesin violence rates among our 37 states that changed their divorcelaws to allow unilateral divorce with 2 alternative control groupsthe 5 states that are yet to adopt unilateral divorce (AR DE MSNY and TN) and the 9 states whose preexisting regime involved


FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de


unilateral divorce (AK LA MD NC OK UT VA VT and WV)18

If there is an underlying relationship between domestic violenceand divorce regime we would expect to observe changing violencepropensities in the treatment group relative to the controls Be-cause the survey universe consists only of couples living in aconjugal unit we are limited to analyzing rates of domestic vio-lence within intact marriages Thus we cannot directly disentan-gle whether the estimated effects reflect a decreasing propensitytoward spousal violence or an increasing propensity for abusedspouses to exit their marriages

Table II analyzes household-level data in which the depen-dent variable Domestic Violence is a dummy indicatingwhether the specified type of violence occurred within each house-hold19 We estimate

Domestic Violenceist Treatments Yeart1985 Treatments

t

t Yeart s

s states controlsiist

where Treatment is a dummy variable that is equal to one if thestate adopted unilateral divorce prior to 1985 and is zero other-wise and is the difference-in-difference estimator

The first row of Table II shows the mean rates of violenceacross households Perhaps surprisingly men are as likely to bephysically abused by their spouses as women are20 The next rowshows difference-in-difference estimates of the effects of unilat-eral divorce on domestic violence Domestic violence towardwomen declined by 17 percentage points in reform states androse 25 percentage points in the control states Thus the differ-ence-in-difference estimate suggests that the treatmentmdashadop-tion of unilateral divorcemdashled domestic violence rates to declineby 43 percentage points or by around one-third over the 1976ndash1985 period Adding state fixed effects in the next row sharpens

18 The 1976 survey did not sample from all states and hence we are not ableto include the following states in our analysis AK AR DC DE HI IA KY MAND NH NM NV RI SD WY

19 The definition of domestic violence follows Straus and Gelles [1994] whocode domestic violence as occurring if there has been any incident over the lastyear in which a person threw something at their partner pushed grabbedshoved slapped kicked bit hit with fist hit or tried to hit with object beat up orthreatened or used a gun or knife against their partner

20 One reason that physical abuse may be perceived as occurring more oftenby men toward women is that assaults by men are seven times more likely toresult in injuries that require medical treatment See Stets and Straus [1990]


these estimates somewhat and these large effects are all found tobe statistically significant The following three rows show thatthese results are robust to the inclusion of a rich set of individual-level controls the set of within-state time-varying economic andsocial policy controls used in Table I and also the use of a probitestimator Further dropping specific states from the sample didnot appreciably change these results

Comparing these declines in violence rates with their baserates domestic violence appears to have declined by somewhere

TABLE IIEFFECTS OF UNILATERAL DIVORCE ON DOMESTIC VIOLENCE

Overall violencea Severe violencea

Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband

Average incidence of each type of violence

117 119 34 46

Estimated change in violence rates in treatment states relative to control states

OLS (Diffs-in-diffs) 43 27 11 29(19) (18) (13) (10)

Add state fixed effects 55 32 20 36(18) (15) (09) (07)

Add individual controlsb 48 19 18 34(17) (14) (10) (09)

Add state-level time-varying 38 18 18 30controlsc (18) (13) (10) (07)

Probit with individual controlsb 47 20 12 21(16) (13) (07) (07)

Sample n1976 2102 n1985 3874 (includes cross-section and state oversamples excludes observations fromstates that are not present in the 1976 data sampling weights are applied) Robust standard errors are in parenthesescorrected for clustering within 72 state-year cells denote significance at the 10 percent 5 percent and 1percent levels respectively All regressions include year fixed effects Dependent variable is a dummy variable set equalto one if the household reports a violent incident as having occurred between spouses over the preceding year and zerootherwise Thus reported coefficients reflect the change in the relevant spousal violence rate in treatment relative tocontrol states in percentage points To assess these changes in percentage terms compare the reported coefficient withthe corresponding term in the first row Each entry reflects a separate regression

a Severe violence is defined as kicked bit hit with fist hit or tried to hit with something beat uppartner threatened with gun or knife or used a gun or knife in the past year Overall violence also includesthrew something at partner pushed grabbed or shoved and slapped (Follows Gelles and Straus [1994])

b Individual controls include a saturated set of dummies for respondentrsquos age race and gender and theeducational attainment and current labor force status of both husband and wife These regressions alsoinclude state-fixed effects

c State-level time-varying controls include the maximum level of AFDC for a family of four in thatstate-year the proportion of the population on welfare the ratio of female to male employment rates thestate unemployment rate and log personal income per capita


between a quarter and a half between 1976 and 1985 in thosestates that reformed their divorce laws We now turn to an alter-native indicator of spousal abusemdashintimate homicidemdashto furtherprobe the robustness of these results

We consider several definitions of intimate homicide Thenarrowest only includes spousal homicide the next group in-cludes homicides committed by any family member or romanticinterest and finally we expand our treatment group to our broad-est categorization which includes all homicides committed bynonstrangers The defect of the broader measures is that thetreatment group is defined to include many relationships that arenot affected by the treatment of unilateral divorce The defect ofnarrower measures is that police classifications of victim-perpe-trator relationships as ldquospousalrdquo are likely to have changed overtime possibly in a way that is correlated with family law regimesleading to (difficult to sign) bias issues21 Further identifyingintimates narrowly such as by ldquospousesrdquo is more likely to sufferfrom endogeneity problems as the legal status that people choosefor their relationships may change with changes in the legalregime

Table III suggests a large and significant decline in intimatefemicide following the adoption of unilateral divorce for all threedefinitions of intimate homicide with column (1) suggesting de-clines on the order of around 10 percent Column (2) shows thatthis estimate is robust to adding a rich set of controls includingnot only the economic social policy and demographic variablespreviously considered but also a set of criminal justice variablesincluding a death penalty indicator Donahue and Levittrsquos Effec-tive Abortion Rate and the share of the statersquos population inprison population rate lagged one year

The results for males murdered are imprecisely estimatedand would admit large effects in either direction The estimateschange substantially across different definitions of intimate ho-micide and adding controls leads to moderate changes in theestimates Dee [2003] has also analyzed these data employing

21 While the coding of married partners as ldquospousesrdquo presents no difficultycoding of common-law marriages cohabiting couples romantic partners andseparated spouses is likely to have changed over time Although these groups maybe small compared with the whole population we do not know if this is true of thehomicidal population All that is known with certainty is that a homicidal memberfrom one of the above groups would not have been coded as a stranger which is themotivation for looking at the broadest of our definitions of the treatment group


count data methods on a short (1968ndash1978) panel He finds alarge increase in males murdered by their spouses22 The sensi-tivity of both sets of results to small changes in specificationmakes us reluctant to draw strong conclusions in either directionfor male homicide

The results for female homicide are more robust and we turn

22 His paper contains a reconciliation of his results with ours which largelyturns on his shorter sample period coding of intimate homicide and functional form

TABLE IIIEFFECT OF UNILATERAL DIVORCE ON INTIMATE HOMICIDE (PERCENT CHANGE)

Nocontrols Including controls

Intimatehomicide

Intimatehomicide

Placebononintimate

homicide

Diffs-in-diffs-in-diffs(intimate lessnonintimate)

(1) (2) (3) (4)

Women murdered by intimates

By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)

Men murdered by intimates

By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)

Sample 1968ndash1994 Sample excludes Illinois due to missing observations from Chicago Police Depart-ment Also excludes Washington DC as an outlier n 1323 Robust standard errors are reported inparentheses and denote significance at the 10 percent 5 percent and 1 percent levels respectivelyDependent variable is the annual intimate homicide rate in each state Each cell reports the estimated effectof unilateral divorce laws from a separate regression The rows focus on different definitions of ldquointimatehomiciderdquo while columns report different specifications Reported coefficients reflect the percentage changein the relevant homicide rate attributed to Unilateral Divorce laws calculated using the unweighted cellmean as the base All regressions include (significant) state and year fixed effects

Controls include an indicator variable for the death penalty the Donahue and Levitt Effective AbortionRate and the state incarceration rate once lagged as well as the AFDC rate for a family of four the natural logof state personal income per capita the unemployment rate the female-to-male employment rate age compositionvariables indicating the share of statesrsquo populations aged 14ndash19 and then ten-year cohorts beginning with age 20up to a variable for 90 and the share of the statersquos population that is Black White and other


to timing evidence to assist us in interpreting these results Aswith the suicide data we once again replace the single dummyvariable Unilateral in the baseline model with several dummyvariables indicating the number of years since (or until) the lawwent (goes) into effect We run this regression for all three cate-gories of intimate homicide The estimated coefficients for fe-males murdered are shown in Figure II For clarity standarderror bands are not shown but as a rough indicator estimatedstandard errors for each lead or lag plotted are around twicethat shown in the corresponding row of Table III

Figure II confirms the initial findings of a decrease in womenmurdered in the period following the passage of divorce lawreforms However the timing evidence is somewhat worryingand the reader is left to judge whether the decline in homicidepredated the law change to an extent that undermines our re-sults This raises the possibility that our regression results maybe picking up the effects of an alternative phenomenon thatpredated divorce law reform

FIGURE IIEffect of Unilateral Divorce on Females Murdered by Intimates

Figure II shows the estimated coefficients (evaluated as elasticities at theunweighted cell means) from three regressions each focusing on a differentdefinition of the female intimate homicide rate Each line plots the coefficients ondummies indicating whether unilateral divorce laws have been in effect for 1ndash2years 3ndash4 years 5ndash6 years etc as shown dummies are also included for similarleads State and year fixed effects are also included


The fact that family law affects behavior between intimatesbut not between strangers provides an opportunity to furtherprobe these results Specifically nonintimate homicide may serveas an ideal placebo group Column (3) of Table III shows thedifferences-in-differences (panel) estimates for the nonintimatehomicide placebo group (that is the dependent variable is theaggregate homicide rate less the relevant definition of intimatehomicide) These results suggest that there is a negative correla-tion between nonintimate homicide and divorce laws albeit not astatistically significant one These results also give us a chance toassess an alternative counterfactual instead of assuming that inthe absence of divorce reform intimate homicide would remainunchanged (as in the first two columns) the differences-in-differ-ences-in-differences in column (4) assumes that the change innonintimate homicide is the relevant baseline These triple-dif-ference estimates suggest that intimate femicide declined whencompared with this counterfactual but that this difference is notstatistically significant (For men the estimates remain bothimprecise and sensitive to changes in definition) Finally othercrime measures provide a further set of interesting placebos andthese results generally show little correlation between statecrime trends and divorce laws These results are reported inStevenson and Wolfers [2003]

VI CONCLUSION

Our analysis examines the effect of unilateral divorce laws onmeasures of extreme marital distress Changes in divorce law ledto one spouse being able to obtain a divorce without his or herpartnerrsquos consent Examining state panel data on suicide domes-tic violence and murder we find a striking decline in femalesuicide and domestic violence rates arising from the advent ofunilateral divorce Total female suicide declined by around 20percent in the long run in states that adopted unilateral divorceWe believe that this decline is a robust and well-identified resultand timing evidence speaks clearly to this interpretation There isno discernible effect on male suicide

Data on conflict resolution reveal large declines in domesticviolence committed by and against both men and women instates that adopted unilateral divorce Furthermore we find sug-gestive evidence of a decline in females murdered by intimatesalthough these results are not as convincing As with suicide


there is no discernible effect on males murdered although thisreflects the imprecision and volatility of our estimates

While our results are open to the interpretation that thelarge declines identified are the result of changing marriage anddivorce rates we believe that this is only part of the story Thetiming evidence suggests that changes in the divorce rate explainlittle of our findings as the long-term effect of unilateral divorceon divorce rates is smaller than the short-term impacts whilesuicide rates show the opposite pattern This suggests importantroles for changes in marital formation and bargaining withinmarriage If unilateral divorce were causing people to make bet-ter matches then this effect would show up slowly over time asthe stock of married people shifted toward those married afterunilateral divorce Beyond this a more complete account musttake changes in marital dynamics into account Unilateral di-vorce changed the distribution of bargaining power within mar-riages and therefore impacted many marriages

Speculating on the policy implications of emerging models ofthe family Lundberg and Pollak [1993 p 992] argued that thepossibility of ldquothe dependence of intrafamily distribution on thewell-being of divorced individuals provides a mechanism throughwhich government policy can affect distribution within mar-riagerdquo The mechanism examined in this paper is a change indivorce regime and we interpret the evidence collected here as anempirical endorsement of the idea that family law provides apotent tool for affecting outcomes within families

WHARTON UNIVERSITY OF PENNSYLVANIA

WHARTON UNIVERSITY OF PENNSYLVANIA AND CEPR IZA AND NBER

REFERENCES

Becker Gary S A Treatise on the Family (Cambridge MA Harvard UniversityPress 1993)

Campbell Jacquelyn C ldquoIf I Canrsquot Have You No One Can Power and Control inHomicide of Female Partnersrdquo in Jill Radford and Diana E H Russell edsFemicide The Politics of Women Killing (New York NY Twayne 1992)

Cutler David M Edward M Glaeser and Karen Norberg ldquoExplaining the Risein Youth Suiciderdquo in Jonathan Gruber ed Risky Behavior Among YouthsAn Economic Analysis (Chicago IL University of Chicago Press 2001)

Dee Thomas S ldquoUntil Death Do You Part The Effects of Unilateral Divorce onSpousal Homicidesrdquo Economic Inquiry XLI (2003) 163ndash183

Fox James A ldquoUniform Crime Reports [United States] Supplementary Homi-cide Reports 1976ndash1994rdquo ICPSR Study 6754 Boston MA NortheasternUniversity College of Criminal Justice 1996

Friedberg Leora ldquoDid Unilateral Divorce Raise Divorce Rates Evidence fromPanel Datardquo American Economic Review LXXXVIII (1998) 608ndash627


Gelles Richard J and Murray A Straus PHYSICAL VIOLENCE IN AMERI-CAN FAMILIES 1985 [Computer file] 3rd ICPSR release Durham NHUniversity of New Hampshire Family Research Laboratory [producer] 1988Ann Arbor MI Inter-university Consortium for Political and Social Research[distributor] 1994

Hamermesh Daniel S and Neal M Soss ldquoAn Economic Theory of SuiciderdquoJournal of Political Economy LXXXII (1974) 83ndash98

Jacob Herbert Silent Revolution The Transformation of Divorce Law in theUnited States (Chicago IL University of Chicago Press 1988)

Lundberg Shelly and Robert A Pollak ldquoSeparate Spheres Bargaining and theMarriage Marketrdquo Journal of Political Economy CI (1993) 988ndash1010

Markowitz Sara ldquoThe Price of Alcohol Wife Abuse and Husband Abuserdquo South-ern Economic Journal LXVII (2000) 279ndash303

McElroy Marjorie B and Mary Jean Horney ldquoNash-Bargained Household Deci-sions Toward a Generalization of the Theory of Demandrdquo InternationalEconomic Review XXII (1981) 333ndash349

Rasul Imran ldquoThe Impact of Divorce Laws on Marriagerdquo mimeo University ofChicago 2004

Riedel Marc and Margaret Zahn ldquoTrends in American Homicide 1968ndash1978Victim-Level Supplementary Homicide Reportsrdquo ICPSR Study 8676 AnnArbor MI Inter-university Consortium for Political and Social Research1994

Stets Jan E and Murray A Straus ldquoGender Differences in Reporting MaritalViolence and its Medical and Psychological Consequencesrdquo in Murray AStraus and Richard J Gelles eds Physical Violence in American FamiliesRisk Factors and Adaptations to Violence in 8145 Families (New BrunswickNJ Transaction Publishers 1990)

Stevenson Betsey and Justin Wolfers ldquoBargaining in the Shadow of the LawDivorce Laws and Family Distressrdquo NBER Working Paper No 10175 2003

Straus Murray and Richard J Gelles PHYSICAL VIOLENCE IN AMERICANFAMILIES 1976 [Computer file] conducted by Murray A Straus Universityof New Hampshire and Richard J Gelles University of Rhode Island 2ndICPSR ed Ann Arbor MI Inter-university Consortium for Political andSocial Research [producer and distributor] 1994

United States Department of Health and Human Services National Center forHealth Statistics ldquoMortality Detail File External Cause Extract 1968ndash19781979ndash1980rdquo ICPSR Study 8224 Ann Arbor MI Inter-university Consortiumfor Political and Social Research 1985

Wolfers Justin ldquoDid Unilateral Divorce Raise Divorce Rates A Reconciliationand New Resultsrdquo American Economic Review (2006) forthcoming


1990s and 2000s Widespread concern over the decline of theAmerican family has led many to point the finger at unilateraldivorce laws claiming that easy access to divorce underminestraditional family structures Unfortunately much of the publicdiscussion centers on the consequences of divorce rather than theconsequences of divorce laws

Unilateral divorce allows marriages to end where one personwants out of the marriage and the other person wants to remainmarried This paper seeks to answer the question who benefitedfrom this change and by how much While models of the familythat rely on a common preference function or internal threatpoints predict little change external threat-point models tell usthat unilateral divorce changes bargaining within marriage byimproving the outside options of each spouse As such bargainingpower and therefore resources shifts toward the person whomost wants out of the marriage

The people most likely to benefit from unilateral divorce aretherefore those who stand to gain the most from having the optionto exit their relationship One possibility is that those in violentpotentially lethal relationships have the most to gain when theycan credibly threaten to exit the relationships Unilateral divorcehas two possible effects on these relationships The first is that itallows them to end2 The second is that the threat of divorce maybe sufficient to prevent future abuse in relationships that con-tinue Our focus in this paper is the effect of allowing unilateraldivorce on these particularly bad marriages potentially throughboth channels

Without access to unilateral divorce people ldquotrappedrdquo in abad marriage had few choices While they could leave the mar-riage without being granted a divorce they would not be able totake any assets from the marriage and would be unable to re-marry We consider the possibility that violent relationships weremore likely to end through suicide or homicide prior to unilateraldivorce Suicides could result from unhappiness the value ofcontinuing to live in the abusive relationship falling below theoption value of staying alive3 Alternatively those in abusiverelationships may have used strategic suicide attempts as a way

2 While fault-based divorce does offer divorce for violent relationships theviolence must be proved in court These cases were quite adversarial and manyabuse victims were likely afraid of the heightened threat during the trial

3 For an economic model of suicide see Hamermesh and Soss [1974]











WITHIN MARRIAGE




































IV SUICIDE RESULTS





Suicide ratest k

kUnilateralstk

s

sStates t









Column no


(1f) (2f) (1m) (2m)

Year of change 16 13 08 14(38) (34) (22) (21)










19 years later 187 164 39 43(32) (33) (20) (20)




97(23)

83(23)

15(13)

20(13)
















FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES































WITHIN MARRIAGE




































IV SUICIDE RESULTS





Suicide ratest k

kUnilateralstk

s

sStates t









Column no


(1f) (2f) (1m) (2m)

Year of change 16 13 08 14(38) (34) (22) (21)










19 years later 187 164 39 43(32) (33) (20) (20)




97(23)

83(23)

15(13)

20(13)
















FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES
























WITHIN MARRIAGE




































IV SUICIDE RESULTS





Suicide ratest k

kUnilateralstk

s

sStates t









Column no


(1f) (2f) (1m) (2m)

Year of change 16 13 08 14(38) (34) (22) (21)










19 years later 187 164 39 43(32) (33) (20) (20)




97(23)

83(23)

15(13)

20(13)
















FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES



















































IV SUICIDE RESULTS





Suicide ratest k

kUnilateralstk

s

sStates t









Column no


(1f) (2f) (1m) (2m)

Year of change 16 13 08 14(38) (34) (22) (21)










19 years later 187 164 39 43(32) (33) (20) (20)




97(23)

83(23)

15(13)

20(13)
















FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES














































IV SUICIDE RESULTS





Suicide ratest k

kUnilateralstk

s

sStates t









Column no


(1f) (2f) (1m) (2m)

Year of change 16 13 08 14(38) (34) (22) (21)










19 years later 187 164 39 43(32) (33) (20) (20)




97(23)

83(23)

15(13)

20(13)
















FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES









































IV SUICIDE RESULTS





Suicide ratest k

kUnilateralstk

s

sStates t









Column no


(1f) (2f) (1m) (2m)

Year of change 16 13 08 14(38) (34) (22) (21)










19 years later 187 164 39 43(32) (33) (20) (20)




97(23)

83(23)

15(13)

20(13)
















FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES
































IV SUICIDE RESULTS





Suicide ratest k

kUnilateralstk

s

sStates t









Column no


(1f) (2f) (1m) (2m)

Year of change 16 13 08 14(38) (34) (22) (21)










19 years later 187 164 39 43(32) (33) (20) (20)




97(23)

83(23)

15(13)

20(13)
















FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES
























IV SUICIDE RESULTS





Suicide ratest k

kUnilateralstk

s

sStates t









Column no


(1f) (2f) (1m) (2m)

Year of change 16 13 08 14(38) (34) (22) (21)










19 years later 187 164 39 43(32) (33) (20) (20)




97(23)

83(23)

15(13)

20(13)
















FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES























Suicide ratest k

kUnilateralstk

s

sStates t









Column no


(1f) (2f) (1m) (2m)

Year of change 16 13 08 14(38) (34) (22) (21)










19 years later 187 164 39 43(32) (33) (20) (20)




97(23)

83(23)

15(13)

20(13)
















FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES























Column no


(1f) (2f) (1m) (2m)

Year of change 16 13 08 14(38) (34) (22) (21)










19 years later 187 164 39 43(32) (33) (20) (20)




97(23)

83(23)

15(13)

20(13)
















FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES































FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES


























FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES






















FIG

UR

EI

Eff

ects

ofU

nil

ater

alD

ivor

ceL

aws

onF

emal

eS

uic

ide

Eac

hpa

nel

repo

rts

resu

lts

from

ase

para

tere

gres

sion

in

clu

din

gal

lcon

trol

sli

sted

inT

able

Ian

dst

ate

and

year

fixe

def

fect

sB

otto

mri

ght

pan

elt

opli

ne

show

sre

sult

sfr

omag

e-ag

greg

ated

regr

essi

onB

otto

mli

ne

sum

sth

ere

sult

sfr

ompr

eced

ing

pan

els

Sca

leis

onth

eri

ght-

han

dsi

de






t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES


























t

t Yeart s












Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES


























Husbandto wife

Wife tohusband

Husbandto wife

Wife tohusband


117 119 34 46























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES

































Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES



























Intimatehomicide

Intimatehomicide

Placebononintimate

homicide


(1) (2) (3) (4)


By spouse 105 126 37 72(59) (60) (35) (69)

By family 89 88 31 56(44) (44) (42) (61)

By known 87 85 01 79(37) (36) (52) (63)


By spouse 123 39 22 109(92) (90) (28) (96)

By family 19 43 13 02(53) (53) (30) (59)

By known 20 50 27 41(31) (31) (43) (52)










VI CONCLUSION









REFERENCES




























VI CONCLUSION









REFERENCES























VI CONCLUSION









REFERENCES



























REFERENCES




































Documents

BARGAINING IN THE SHADOW OF THE LAW: DIVORCE LAWS …users.nber.org/~jwolfers/Papers/bargaining_in_the_shadow_of_the_law.pdfunilateral divorce in California. This legislative change