From 1999 to 2013,U.S. mortgage debt doubled and then contracted sharply. Our understanding of the factors driving this volatility in the stock of debt is hampered by a lack of data on mortgage flows. Using comprehensive, individual-level panel data on consumer liabilities, I estimate detailed mortgage inflows and outflows. During the boom, inflows from real estate investors tripled, far outpacing growth from other segments such as first-time homebuyers. During the bust, although defaults and deleveraging are popular explanations for the debt decline, a collapse in inflows has been the major driver. Inflow declines across counties have been associated not just with house price declines, but also with rising unemployment and higher minority population shares. Finally, inflow declines reflect, in part, a dramatic decline in first-time homebuying. First-time homebuying fell among both high and low credit score individuals, but much more precipitously for low score individuals. Further analysis suggests that the differential decline by credit score likely reflects markedly tightened credit supply. This paper updates and extends an earlier paper titled, “Mortgage Debt and Household Deleveraging: Accounting for the Decline in Mortgage Debt using Consumer Credit Record Data.” Federal Reserve Board, 20thand C St NW, Washington, DC. neil.bhutta@frb.gov. The views and conclusions expressed in this paper represent those of the author, and do not necessarily reflect those of the Federal Reserve Board or System.
These swings in the stock of debt are well known from standard data sources, but there is virtually no data on the underlying flows, resulting in a limited understanding of the reasons for the recent volatility, and hindering the ability of policy makers and prudential regulators to know whether and how to react to fluctuations in aggregate credit growth. This paper aims to help fill this gap by developing novel measures of mortgage flows from comprehensive, individual-level panel data on consumer liabilities. At the highest level, mortgage inflows come from consumers who accumulate mortgage debt over a short period, and outflows come from consumers who reduce such debt. More importantly, I also disaggregate these broad inflows and outflows into key subcomponents according to several borrower characteristics (e.g. credit score) and activities (first-time homebuying, default, etc.) During the housing boom, a number off actors could have contributed to the buildup of mortgage debt such as equity extraction (e.g. Greenspan and Kennedy 2008; Mian and Sufi 2011; Bhatta and Keys 2013), increased purchases of second homes and investment properties (e.g. Haughwaut et al. 2011), and increased first-time homebuying by more marginal borrowers due to a credit expansion that may have been partially government induced (e.g. Mian and Sufi 2009; Agarwal et al. 2012).