The dominant role of residential real estate investment in the feeding of the 2000s expansion is universally recognized. Its tepid performance in the current expansion is also conspicuous, shown in Graph 1. The set of explanations for the uninspiring growth center around credit availability, employment and incomes, the student debt burden, household formation rates, and shifts in preference with respect to homeownership (Larrimore, Schuetz, and Dodini 2016). While these are vital elements of the tableau, some further perspective is recommended to help clarify our understanding of the current business cycle. Perhaps we should, like Furiosa in Mad Max: Fury Road, go back to get ahead. For example, recall that the theme of the 2007 Kansas City Fed’s annual Jackson, WY Symposium was “Housing, Housing Finance, and Monetary Policy”, and included a presentation titled “Housing Is the Business Cycle” (Leamer 2007). And that home price appreciation was so pronounced and established in the Noughts that the wealth effect from housing was compelling subject matter; see, e.g., Case, Quigley and Shiller (2006). Between then and now has been deleverage. We present here a series of graphs of US public and private debt over the last quarter-century, the aim to give context to leverage trends since the late 2000s crisis.
Deleveraging – when (parts of) the transmission mechanism of monetary policy works in reverse – is typically a time-consuming process. It can slow recovery and destroy potential through a number of different channels. It has the capacity to affect the profits and capital of businesses, and incomes and wealth of households, directly and indirectly. Financial repression can take hold, significantly so in some cases; often, there is a moral dimension that adds heft to the repression. The drag from excessive leverage emits sounds about the expected vigor of an emerging recovery, potentially draining confidence.
Given the nature and depth of the 2007-2009 crisis, it was clear at the time that significant deleveraging among consumers was set to happen; that is, considering the volume of badly underwritten mortgages that were originated in the 2000s, a pickup in defaults, short sales and paydowns was bound to result out of an environment featuring declining home prices, tight credit, and sagging domestic and global economies. Perhaps other classes of debt would be caught up as well. Given that the need to delever was well-understood by policymakers, it is natural to ask how well the transition was managed. Our goal is to make an assessment over the course of a few blog posts, but first we should understand what the transition looks like, and whether it has concluded.
In the wake of the crisis, deleveraging has been accompanied by massive and still-ongoing banking reform as well as unconventional monetary policy. Because it was largely funded with (excess) banking reserves, quantitative easing abetted the forced deleveraging of large banks. These policy actions have shaped the nature of the recovery. And while the process of end-user deleveraging is not insensitive to the level of interest rates, the stimulus from low rates is stunted because of diminished demand for new credit. The slow process of banking reform along with tight underwriting standards also weakens the stimulus delivered by low interest rates. It is important to keep in mind that credit suppression is faced by all potential borrowers.
Richard Koo (e.g., 2014) has been in the modern-day vanguard of economists who have explored deleveraging, Koo noting, i.a., its unfortunate necessity, the limitations of monetary policy efficacy in such episodes, and the benefits of public sector stimulation of income growth in order to hasten the paydown of debt. Still, just as the full macroeconomic impact and potential risks of leverage are incompletely understood — see, for example, Jonathan Ostry, Atish Ghosh and Raphael Espinoza (2015), who note with respect to sovereign debt: “What constitutes a safe level of debt…is, needless to say, very difficult to pin down precisely in practice, and can never be established through some mechanical rule or threshold” — so as well for the impact and risks of deleveraging.
Yet it takes only a glance at the data to see that the monstrous expansion and subsequent contraction of residential mortgages over the last 15 years might have had a lasting, material effect on employment, incomes, consumption, and wealth. So, it is worthwhile to check in periodically on levels of outstanding debt. John Geanakoplos and Lasse H. Pedersen (2011) express similar views about the value of simple, straightforward ongoing monitoring of leverage, albeit applying different measures than the ones we have used here, such as margin requirements, and especially with respect to recent originations.
The graphs below, using data from the Federal Reserve Board’s Financial Accounts of the United States, show outstanding debt for households/nonprofits, nonfinancial businesses, and government, scaled to National Income (NI). We use NI and not GDP as NI excludes the consumption of public and private fixed capital: for Q4 2015, NI clocked in at USD 15,698 billion saar; the sum of private and public capital consumption and a statistical discrepancy totaled USD 2,627 billion. NI also includes net income from abroad (USD 195 billion). NI is useful to apply as a common base across categories, though its relevance as an indicator of ability-to-pay is limited.
It should be kept in mind that, although we have not run any statistical tests, debt growth is expected to lag income. This would be more apparent at inflection points. Thus we see private debt/NI continuing to rise into the last two recessions. John Lonski (2016) likewise notes that changes in levels of corporate debt growth lag changes in profit growth. This has implications for employment and capital expenditures especially in early stages of recovery; this has been noted by, among others, Satyajit Chatterjee (2013).
Graph 2 shows the ratio of total outstanding debt of households/nonprofits, nonfinancial businesses, and the federal government to NI. State and local government debt was omitted in this graph in order to get a clearer view of the extraordinary expansion of federal government debt in the wake of the Noughts Panic, but will be examined further on. Also omitted is debt of financial businesses, which is difficult to interpret given unknown multiple layers of intermediation; this is highly unfortunate given that financial businesses occupy much of the core of this discussion!
The next four graphs plot the year-over-year change in debt outstanding of household mortgages, consumer credit, nonfinancial businesses, and government, as a ratio to the year-over-year change in NI. Using quarterly observations over Q1 1990 to Q4 2015, change is defined as valuet+4/valuet. Admittedly, this is an uncommon measure of debt, chosen to give greater definition to the ebbs and flows of all categories of debt relative to income. Note the graphs do not share a common range.
Graph 3 examines home mortgages alongside total consumer credit. Graph 4 breaks out the major categories of consumer credit: student loans, autos, and credit cards. Mortgage outstandings shrank unrelentingly for six years: from a prior peak of USD 10,694 billion in Q1 2008, total home mortgage debt outstanding fell 12.4 percent to a low of USD 9,370 billion in Q2 2014. But in the six quarters since the low (marking against the most recent observation, Q4 2015), mortgage debt has increased by a feeble 1.3 percent. Other important considerations in play here include the evolution (and devolution!) of funding markets, housing prices, home ownership rates, and household formation.
Credit card debt was also subject to fierce deleveraging in the recession and its aftermath, despite its growth appearing rather tame for much of the Noughts, reflective of underwriting standards. Perhaps card debt got caught in the gears of the mortgage fiasco; perhaps shifts in preferences are reflected. As general purpose debt, the use of credit cards proceeds is something of a mystery. Helu Jiang and Juan M. Sanchez (2016) note an age bias in credit card levering in the period 2004-2008 — nearly 50 percent of the growth attributed to individuals 56 and older — and delevering in the period 2008-2015 — 68 percent of decline in outstandings came from individuals younger than 46, virtually none from those 56 and older.
The recent growth of student loans is discouraging. The history of other series suggests that prolonged growth of five-plus percent over NI growth could spell trouble. But perhaps student loan debt will continue to carve out its own path, perhaps at the expense of other types of debt — and perhaps that only temporarily, as suggested by Alvaro Mezza, Kamila Sommer, and Shane Sherlund (2014) — if significant income disparities between college graduates and others persist.
Graph 5 breaks down nonfinancial business debt (loans and securities) into corporate and noncorporate segments. Corporate debt outstanding is currently running around 1.7 times its noncorporate counterpart. Note that noncorporate business debt grew at a faster pace than corporate debt from 1997 through 2011. According to the Financial Accounts of the United States Tables L.103 and L.104 (release of March 10, 2016), 71 percent of noncorporate business debt stemmed from mortgages as of Q4 2015, compared to 6.5 percent for corporate businesses. Mark E. Schweitzer and Scott A. Shane (2010) found that self-employed households were more likely to carry home equity debt, and with higher median balances, than households working for someone else. The recent flows in noncorporate debt, credit cards, and home mortgages suggest a careful examination of self-employment and small business formation in the current recovery is in order.
Although much has been written about apparent rapid growth of corporate nonfinancial debt in the recovery, corporate debt growth has lagged National Income growth since the Q2 2009 trough: as of Q4 2015, corporate debt outstanding has increased 25 percent since Q2 2009 versus 31 percent for National Income. Of course, confidence in the manageability of the debt burden should be predicated on confidence in the quality of such things as income and the use of proceeds.
Graph 6 shows federal and state/local government debt. We cannot possibly do justice here to the question as to why, six and a half years into the recovery, US federal government debt was at a cyclical high: USD 15,166 billion in Q4 2015. Nor will we examine here the reasons why state and local government debt receded relative to NI. These are critical developments that we hope to address in the near future.
Because absolute magnitudes also matter, we conclude with a table of debt outstanding and NI for the cyclical peaks Q3 1990, Q1 2001, and Q4 2007, and for the most recent observation, Q4 2015. The far right column shows the growth factor for each category from 1990 through 2015.
US Debt Outstanding at cycle peaks and most recent observation, excluding financial businesses and “rest of world”
Case, Karl E., John M. Quigley, and Robert J. Shiller. 2006. Comparing wealth effects: the stock market vs. the housing market. Cowles Foundation Paper No. 1181. http://cowles.econ.yale.edu/
Chatterjee, Satyajit. 2013. Why recovery from a financial crisis can be slow. Federal Reserve Bank of Philadelphia Business Review, Q2. https://www.philadelphiafed.org/research-and-data/publications/business-review/2013
Geanakoplos, John and Lasse H. Pedersen. 2011. Monitoring leverage. Cowles Foundation Paper No. 1838. http://cowles.econ.yale.edu/
Jiang, Helu, and Juan M. Sanchez. 2016. The deleveraging of US households: credit card debt over the lifecycle. Federal Reserve Bank of St. Louis Economic Synopses, Number 11. https://research.stlouisfed.org/publications/economic-synopses/2016/05/27/the-deleveraging-of-u-s-households-credit-card-debt-over-the-lifecycle/
Koo, Richard C. 2014. Balance sheet recession is the reason for secular stagnation. In Secular Stagnation: Facts, Causes and Cures, ed. Coen Teulings, Richard Baldwin. CEPR Press, http://voxeu.org/sites/default/files/Vox_secular_stagnation.pdf
Larrimore, Jeff, Jenny Schuetz, and Samuel Dodini. 2016. What are the perceived barriers to homeownership for young adults? Board of Governors of the Federal Reserve System Finance and Economics Discussion Series 2016-021. http://dx.doi.org/10.17016/FEDS.2016.021
Leamer, Edward E. 2007. Housing is the business cycle. Federal Reserve Bank of Kansas City Symposium Proceedings. https://www.kansascityfed.org/publicat/sympos/2007/PDF/Leamer_0415.pdf
Lonski, John. 2016. The three stages of the credit cycle: Stage III. Moody’s Analytics Credit Markets Review and Outlook, May 12. http://www.moodys.com/
Mezza, Alvaro, Kamila Summer, and Shane Sherlund. 2014. Student loans and homeownership trends. Board of Governors of the Federal Reserve System, FEDS Notes. October 15. http://www.federalreserve.gov/econresdata/notes/feds-notes/2014/student-loans-and-homeownership-trends-20141015.html
Ostry, Jonathan D., Atish R. Ghosh, and Raphael Espinoza. 2015. When should public debt be reduced? International Monetary Fund Staff Discussion Papers SDN/15/10, June. http://www.imf.org/external/pubs/ft/sdn/2015/sdn1510.pdf
Schweitzer, Mark E. and Scott A. Shane. 2010. The effect of falling home prices on small business borrowing. Federal Reserve Bank of Cleveland Economic Commentary, No. 2010-18, December 20. https://www.clevelandfed.org/newsroom-and-events/publications/economic-commentary/economic-commentary-archives/2010-economic-commentaries/ec-201018-the-effect-of-falling-home-prices-on-small-business-borrowing.aspx