“…Existing research shows that demand for credit card loans and other forms of unsecured debt is countercyclical and increases when households become more liquidity constrained (Drozd and Kowalik, 2018;Collins, Edwards, and Schmeiser, 2015;Sullivan, 2008;Bird, Hagstrom, and Wild, 1999). Instead of a bias away from zero, the countercyclical relationship creates a bias toward zero for the interaction between job loss and credit card loans, and we view our estimate of β 2 in Equation 3 as a conservative lower bound of credit card loans' moderating effect on college-persistence rates.…”