In 2012, there were over 1 million personal bankruptcy filings and over 1 million divorce filings in the U.S.  The connection between bankruptcies and divorces is well-documented.  Surveys of bankruptcy filers routinely list divorce among the three most common reasons for filing.  Filing for bankruptcy entails mandatory credit counseling, which may help couples improve their finances and lessen the probability of divorce.  Filing for divorce generates legal costs and property division debts, some of which can be eliminated in bankruptcy.  Despite this intuitive relationship, the presence of financial stress as a critical third factor affecting both divorce and bankruptcy makes causal effects difficult to estimate.  In this paper, I examine the causal relationship between filing for divorce and filing for bankruptcy.

            Both divorces and bankruptcies are major financial and social events for individuals that generate externalities relevant to policymakers.  Personal bankruptcies impose costs on creditors, who respond by increasing interest rates and restricting the supply of credit to low asset households to minimize risk.  Divorces lead to burdensome legal bills, loss of household economies of scale, and negative impacts on children.  These harmful spillovers have inspired numerous policy efforts to lower both the divorce and bankruptcy filing rates.  Understanding the causal link between these events reveals how policies designed to affect bankruptcy may influence divorce or the reverse, as well as improving predictions of future bankruptcy and divorce rates and the resulting economic and social consequences.

            Recent research on the effects of adverse events on bankruptcy filings finds that individuals who get divorced have a higher probability of filing for bankruptcy both before and after divorce than those who never divorce.  This pattern could result from divorces causing bankruptcies, bankruptcies causing divorces, financial stress or another unobserved factor causing both, or a combination of all these explanations.  To resolve this problem, I identify two plausibly exogenous law changes that shift the timing of divorces and bankruptcies.  The 2005 bankruptcy reform was designed to make bankruptcy less generous to debtors, causing a large increase in bankruptcy filings as individuals sought to take advantage of the pre-reform law.  Similarly, the passage of state unilateral divorce laws in the 1960s and 1970s caused temporary increases in divorce filings.  I use variation from predicted aggregate divorce and bankruptcy filings in the year immediately before or after the law change to identify whether the frequency of one event affects the frequency of the other.

            I collect data on state and county divorce rates from state vital statistics offices and district and county bankruptcy filings from federal judicial business reports, covering 1940-2012 in an unbalanced panel that includes the vast majority of U.S. divorce and bankruptcy filings.  I use business bankruptcy filing data as a falsification test.  Other county and state data come from the County and City Data Book and the Bureau of Economic Analysis Regional Economic Information System.  I test robustness using seven different sets of unilateral divorce law passage dates from the law, economics, and sociology literatures.