Manipulating Random Assignment : Evidence From Consumer Bankruptcies in the Nation's Largest Cities

Random case assignment is thought to be an important feature of decision-making in federal courts because it helps guard against favoritism (actual or perceived) toward particular parties or types of cases. In bankruptcy courts, cases are randomly assigned to both judges and trustees. In Chapter 7 cases, for example, the trustee is a quasi-judicial actor, typically a private-sector lawyer, who has been selected to audit the debtor's finances, find and liquidate assets, and police compliance with the law. We study three major bankruptcy jurisdictions (covering Chicago, Los Angeles, and parts of New York) and find that the random-assignment process for Chapter 7 trustees is failing in two of them (Chicago and New York). We introduce several measures of non-random assignment. Across all measures, random assignment is failing: Trustees within the same court have substantially different case characteristics, despite the purportedly random assignment process. We present evidence that the imbalance in case characteristics is caused by attorney manipulation: Attorneys strategically time their case filings to avoid or attract particular trustees ("trustee shopping"). By contrast, among cases filed by debtors who have not hired attorneys ("pro se filers"), there is no case imbalance across trustees. Because they do not engage in manipulation, pro se filers – who account for the bottom decile of income and asset values among Chapter 7 debtors – are the debtors most burdened by trustee-shopping by bankruptcy attorneys. We conclude by presenting evidence that trustee-shopping is less prevalent in Los Angeles due to differences in its random-assignment protocol