A Focus on 'Systemic Risk' Is Unlikely To End Well
Earnest proclamations by politicians and regulators about systemic risk have been de rigueur since the 2007-09 global financial crisis. The 2010 Dodd-Frank Act created many new regulations and two new agencies to identify and act upon the alleged danger. The Office for Financial Research (OFR), housed at the U.S. Treasury, is supposed to collect and provide the data on potential systemic risks, and the Financial Stability Oversight Council (FSOC), chaired by the Secretary of the Treasury, is charged to act upon the information before they become a threat to financial stability.
What drives systemic risk? One theory sees it as the result of "domino effects" arising within the banking system. If losses lead banks to de-lever and sell off risky assets, the sudden dumping of these assets could cause their prices to fall, endangering the solvency of other institutions. Furthermore, if banks are linked to each other via interbank debts, interest rate swaps, and insurance contracts, then the failure of one institution can potentially bring down others as the failed institution is unable to make good on its contractual agreements.
But as much academic research has shown, including an authoritative new book by Xavier Freixas, Luc Laeven, and Jose-Luis Peydro, Systemic Risk, Crises, and Macroprudential Regulation (MIT Press), these potential concerns have not proved to be the main problem. The primary source of trouble has been a combination of high leverage and undiversified risk-taking by banks and other intermediaries who rely on government-insured debt funding.
Two recent NBER working papers by Oscar Jorda, Moritz Schularick and Alan Taylor show that the majority of financial crises over the past several decades have been a consequence of the combination of the rising leverage of banks and their increasing concentration in real estate lending. It is not rocket science to understand that a financial system whose intermediaries use short-term debt to finance risky real-estate assets -- assets with highly correlated risks, that move in sync with the business cycle, and which are hard to liquidate during times when prices fall -- is especially vulnerable to systemic insolvency.
Why would the market let banks structure themselves this way? It wouldn't, if the market was deciding on their structure. But it isn't. A combination of subsidized real estate funding policies - in the U.S. in the recent past, through mortgage-holding Government Sponsored Enterprises (GSEs), the Federal Housing Administration (FHA), the Federal Home Loan Bank (FHLB) System, the Community Reinvestment Act (CRA) - and the insulation of debtholders from default risks through deposit insurance, and bailouts of banks, GSEs and others-remove the market discipline that would otherwise constrain systemic risk.
As Sophia Chen and I are finding in our new empirical study, the dramatic world-wide increase in both banking system leverage and asset risk that has occurred since the 1970s is closely related to the generosity of a country's deposit insurance system. The more generous the deposit insurance protection, the greater the asset risk and the greater the bank leverage. We also find that, in developed economies, the expansion of deposit insurance also leads to greater consumer mortgage lending.
Not only has risky mortgage lending produced instability around the world, it also has crowded out more productive uses of funds. A recent OECD study by Boris Cournede and Oliver Denk ("Finance and Growth in OECD and G20 Countries") shows that in developed economies recent growth in bank credit has been associated with negative economic growth, a reversal of the typical positive link observed between bank credit and growth. The negative effect of credit in the OECD study is driven by the government subsidization of housing finance (largely mortgage finance), which not only has made financial systems vulnerable to collapse, it also crowds out the financing of other growth-producing investments.
Historically, banks had to meet market discipline imposed by depositors, who were unwilling to put money into a bank that created a significant risk of loss for its depositors. Banks were not protected by deposit insurance, nor were they encouraged to make risky mortgage loans. Indeed, U.S. nationally chartered banks prior to 1913 were prohibited from holding any real estate loans. But in 1913, to achieve political support for the establishment of the Federal Reserve System from agricultural areas, the prohibition on real estate lending was relaxed. Over time, real estate lending - once considered incompatible with deposit financing because of the illiquidity of real-estate and its pro-cyclical pricing - became the primary activity of depository institutions, and government has gone from prohibiting real estate lending to encouraging it aggressively via a combination of policies: government protection of banks and GSEs, FHA and FHLB subsidies, and mandates to increase risk imposed on banks via the CRA and on GSEs by the GSE Act of 1992.
Prudential regulations on banks' cash holdings also were relaxed because enforcing them would only have put sand into the gears of the push to expand real estate risk. National banks historically faced very strict requirements on their holdings of cash assets. Money center banks had to hold 25% of their deposits in hard cash. This was meant not only to preserve the health of money center banks; requiring money center banks to have huge cash holdings was intended to stabilize other banks that relied on them for cash, and thereby bolster the system as a whole.
But regulatory and market discipline on banks was put aside in the interest of the political push for subsidized real estate credit. Since the 1980s, as banks' holdings of real-estate loans and related securities grew dramatically, their cash assets (defined broadly as currency, reserves at the Fed, government bonds, and agency bonds) fell dramatically. Broadly defined cash assets for weekly reporting (large) Federal Reserve member banks fell from 20% of assets in 1987 to only 13.5% of assets by January 2008. Real-estate loans for these banks rose from 20% of assets to 33% over that same period, and that rising loan exposure does not take into account the enormous boom in banks' holdings of mortgage-backed securities, on or off their balance sheets, which played such an important role in the recent crisis.
Given that the United States has experienced two major real-estate-caused financial crises in the last three decades (the crisis involving banks and S&Ls in the 1980s, and the recent subprime crisis), and given that we have established a FSOC and OFR to identify systemic risks and prevent the recurrence of such crises, can we expect regulators and politicians to take the necessary actions to avoid another real estate-finance crisis in the future?
Regulatory standards now require higher holdings of cash and reduced leverage, but if the riskiness of lending is sufficiently high, those safeguards will prove inadequate. Unfortunately, Dodd-Frank did nothing to roll back real estate finance subsidies or the government protection of banks and GSEs-indeed, GSE debts were exempted from its ban on bank proprietary trading, and Title II of Dodd-Frank institutionalized bailouts by writing a road map of how they will occur, and determining how they will be funded (by "fees," which are known to everyone outside of Washington as taxes), rather than preventing them.
There has been no progress since then in winding down the GSEs or placing a credible ceiling on banks' credit risk exposures to real estate lending. The Obama administration's appointment of Mel Watt to head the Federal Housing Finance Agency (FHFA) immediately produced a reduction in minimum down payment requirements for GSE mortgages from 5% to 3%, as well as a reduction in FHA insurance premiums.
The result, measured by the American Enterprise Institute's Mortgage Risk Index, has been a rise in risky mortgages. Already, as of February 2016, 53.3% of the mortgages of first-time buyers (who account for 56.7% of mortgages) are "high-risk" mortgages. Over the past two years (since Mel Watt's relaxation of GSE lending standards), AEI's indices measuring their mortgage risk on first-time homebuyers' mortgages has risen from 14 to 15.8, an increase of 13 percent. Loose credit is helping to drive house prices higher, which is adding to the political pressures to further loosen mortgage underwriting standards.
The FSOC and OFR - which are part of the same administration that appointed Mr. Watt - do not point to these developments as causes for concern. They are too busy checking all the new boxes created by the Dodd-Frank Act. The failure of the OFR, the FSOC, the Fed, and everyone else in official Washington to recognize that mortgage risk is rising again, or to try to do anything to limit it, is no surprise. Systemic risk from real estate lending is the product of a political game that often produces bureaucracies whose job is either to magnify it or to create thousands of pages of reports and massive databases that unwittingly serve to distract attention from it. This is not likely to end well.