Household bankruptcies in America increased steeply between 1980 and the mid-2000s, reaching an annual rate of 1.6 million filings in 2004 – representing 1.4% of all US households. In response to this trend, and because of substantial lobbying by the consumer credit industry, Congress passed the Bankruptcy Abuse Prevention and Consumer Protection Act in 2005. The law was intended to prevent ‘abuse’ of the system by wealthy debtors by imposing a means test, mandatory credit counselling, and other new requirements on bankruptcy filers. These changes increased the costs of filing by 60%, from $921 to $1,477 (GAO 2008).
The Bankruptcy Abuse Act succeeded in reducing the number of bankruptcy filings. However, a number of studies have shown that the average income of filers increased after 2005 (Lawless et al 2008, Lindblad et al 2011), casting doubt over its effectiveness in curtailing abuse by wealthier debtors. The stricter filing requirements were also ill timed; the housing market collapse and economic recession likely increased the number of bankruptcies due to contingencies beyond a household’s control. Many Americans seeking bankruptcy relief as a result of foreclosure or job loss are now forced to pay the steep costs targeted at preventing abuse by the wealthy.
Why do tax rebates spur bankruptcies?
In a recent NBER working paper, my co-authors and I document that legal and administrative costs inhibit a significant number of households from filing for bankruptcy (Gross et al 2012). Moreover, the number of households facing these barriers has doubled during this decade. In the paper, we examine how household bankruptcy rates responded to the 2001 and 2008 income tax rebates, using the fact that the rebate payments were distributed randomly based on filers’ social security numbers. We find that after receiving tax rebates ranging from $300 to $1200, households were more likely to file for bankruptcy.1
The positive response of bankruptcy rates to tax rebates suggests that households face liquidity constraints – impediments to purchasing services that benefit them over the long run due to costs today. The extra income from tax rebates allowed previously-constrained households to file for bankruptcy, explaining the increase in bankruptcy rates we observed. We estimate that 2% of households filing for bankruptcy during the rebate period in 2001 faced liquidity constraints prior to receiving rebate cheques, and 4% of filers were constrained during the rebate period in 2008.
Using today’s bankruptcy rates, our results predict that more than 30,000 to 60,000 households will have difficulty saving up for filing costs in 2012. We interpret these estimates as a lower bound on the true number of liquidity-constrained households, because our results consider only households with constraints so severe that they would have been unable to file without the tax rebates. Furthermore, some may have been so constrained that they could not afford bankruptcy even with the rebates.
Bankruptcy policy needs to address the root causes of household debt
Our study documents a significant downside to the Bankruptcy Act’s approach of raising the barriers to bankruptcy. Liquidity-constrained households are likely to have the most to gain from bankruptcy, yet they are the ones screened out by high fees. Moreover, the increased costs do little to mitigate strategic behaviour such as OJ Simpson’s notorious purchase of an expensive home in Florida to exploit that state’s generous bankruptcy provisions.
Figure 1 illustrates the intuition behind how higher fees affect the distribution of bankruptcy filers, based on the simple model in our paper. When filing for bankruptcy costs c, two thresholds A and B define how households with different levels of wealth make their bankruptcy decisions. Those with wealth below threshold A become liquidity-constrained (they would like to file for bankruptcy but cannot afford the fees). Those between thresholds A and B file for bankruptcy, and those with wealth above threshold B are wealthy enough that they prefer to pay their debts rather than go bankrupt. The thresholds are defined by bankruptcy costs c, household debt B, and bankruptcy exemptions e.2
Figure 1. Bankruptcy costs and the distribution of filers
Raising bankruptcy fees – as the Bankruptcy Act did in 2005 – lowers threshold B and raises threshold A. The net result is to reduce the overall number of bankruptcies, but doing so both by inhibiting relatively-wealthier households and by constraining the poorest. Lawmakers did not seem to have taken into account the latter effect when enacting the Bankruptcy Act. While the Act focused on moral hazard as the culprit for rising bankruptcy rates, it failed to take into account the explosion of consumer credit in America since the 1980s. Mian and Sufi (2011) report that the household debt-to-income ratio more than doubled from 0.9 in 1980 to 2.0 in 2009. Against the backdrop of this dramatic rise in household debt, raising the costs of filing is an ineffective strategy for curtailing consumer bankruptcy. The recession has caught many households in a rising tide of unemployment and foreclosure, and high bankruptcy fees prevent them from obtaining much-needed relief.
There are many reasons to be troubled by today’s high bankruptcy rates – more than 1.3% of all US households filed in 2011. But we can only fix America’s bankruptcy problem by eliminating excessive consumer credit, not by adding insult to injury for households that are already broke.
1 Johnson, Parker, and Souleles (2006) and Parker, Souleles, Johnson, and McClelland (2011) have documented that both durable and nondurable household consumption increase significantly after receiving tax rebates.
2 Exemptions are rules that allow households to keep a portion of their assets such as primary residences and vehicles for personal use. These assets are retained by the household and not liquidated or transferred to creditors in bankruptcy. Threshold B is derived by comparing the household’s wealth after filing for bankruptcy (e · (W – c)) with its wealth upon repaying its debt in full (W – B). Further details of the model are available in our paper.