Can serious financial disruptions arise without leverage?

Everyone knows that leverage is a key driver of financial fragility. Since the crisis that began in 2007, regulators around the world have focused on limiting leverage to contain systemic risk, prompting many banks to raise capital and shrink their balance sheets.

Will that suffice to avoid another bout of financial instability? Put differently, is high leverage a necessary condition for a crisis? Some central bankers seem to think so. To secure the financial system, they have come to rely on “macroprudential” tools that limit leverage (e.g. capital requirements and loan-to-value caps), while reserving interest rate policy for the traditional goals of price and economic stability.

Yet, recent analysis casts doubt on this “separation principle” for policy tools. Last year, for example, Governor Jeremy Stein of the Federal Reserve Board highlighted how difficult it is for regulators and supervisors to observe and address the varied ways in which intermediaries take risks that can become systemic. Basically, financial firms have big incentives to avoid (or evade) regulatory constraints. Governor Stein concluded famously that interest rate policy “gets in all of the cracks” that prudential policy might miss.

Recently, the 2014 U.S. Monetary Policy Forum (USMPF) report (which one of us co-authored) went a step further. The report described how the behavior of intermediaries and their investors can trigger “market tantrums” even in the absence of leverage. It also showed that Federal Reserve monetary policy that aims at stimulating the economy can foster such disruptions. And, how the macroprudential tools normally used to contain leverage are not available to address this source of systemic risk should it arise.

To understand the USMPF’s argument, imagine that investors pay intermediaries to manage their funds. Suppose further that these asset managers are sensitive to their relative performance, possibly because they face a penalty for coming in behind their peers. In such an environment, asset managers have an incentive to chase higher returns in good times, selling in unison when a sufficiently adverse shock hits. Such herding creates periods of low market volatility that are eventually (inevitably) followed by an abrupt snapback to high volatility – a market tantrum. The USMPF report provides evidence that changes in inflows to and outflows from bond mutual funds create exactly these sorts of large and sudden shifts in returns.

Importantly, monetary policy itself can be a source of this instability. To see how, consider the forward guidance that some central banks now use to stimulate their economies at the zero interest rate bound. Forward guidance aims to lower long-term interest rates and promote investment.  It works in part by making investors believe that the central bank will keep interest rates low for an extended period, thereby depressing market volatility and risk.

But forward guidance also can build up financial and economic risks for the future. When the herd comes to understand that the central bank will raise interest rates (as it eventually must), the managers’ incentive is to sell – all at the same time – triggering a market tantrum that can weaken the economy more substantially than policymakers expect or wish. The point is that such snapbacks are difficult to control and can defeat policymakers’ initial objective of driving up aggregate demand in order to stabilize inflation and employment.

What to conclude? First, central banks may sometimes wish to use interest rate policy – rather than prudential safeguards – to prevent financial instability. Second, the choices central bankers face are even more complex than is commonly believed. In particular, the tools that they use to keep inflation close to target and unemployment close to a sustainable level today can create risks of financial instability tomorrow. Third, high leverage may be sufficient to destabilize the financial system, but it is not necessary.  As the 2014 US Monetary Forum Report suggests, the answer to the title question is “yes.”