“Is this a Lehman moment?”
In the days after the U.K. Brexit referendum, that was the leading question many people were asking. It is the right question. Unfortunately, despite years of regulatory reform in the aftermath of the financial crisis, the answer is: we don’t know. That is why policymakers are especially worried about heightened financial volatility in the aftermath of U.K. voters’ decision to leave the European Union.
Many people are puzzled by the complexities of Brexit and by the breadth of its financial impact across the globe. They are probably not surprised by the downgrading of U.K. sovereign credit. They may not be surprised by the two-day nose-dive of the U.K. pound (which set a 31-year low versus the U.S. dollar) or the London stock market (which gave up about 6 percent before recovering last week). But, over the same two days, some European stock markets suffered even more than those in London (with the broad indexes in France and Germany initially down by around 10 percent and in Italy by more than 15 percent), while sovereign bond yields in several markets sank to (or close to) record lows. Broadly speaking, market developments ranging from the euro-yen exchange rate to commodities signaled a decline in investors’ tolerance for risk amid heightened global uncertainty and diminished growth prospects.
But has Brexit substantially raised systemic risk? Has it made more likely a scenario in which banks and other financial intermediaries lose trust in one another to the point where credit might cease to flow? Is this a “Lehman moment” where financial institutions’ doubts about other intermediaries’ well-being (or about the reliability of their collateral) will cause short-term funding to dry up, triggering a vicious cycle of fire sales, plunging asset prices, falling capital, broadening financial dysfunction, and recession? Or, less dramatically, but still painfully, will intermediaries’ elevated demand for liquidity limit the supply of credit to healthy borrowers, putting yet another brake on growth?
To answer these questions quantitatively, we would need to know a great deal about the current capitalization of intermediaries—especially those in the United Kingdom and continental Europe—and about their exposure to Brexit risks. From this perspective, it is mildly comforting that 30 (out of 33) large intermediaries passed the Fed’s 2016 stress test without reservations this week, signaling that capital in the U.S. banking system has risen substantially since the crisis of 2007-2009. Over time, the Fed’s annual stress tests have become progressively more rigorous, and now serve as the de facto constraint on capital planning for large U.S. banks. Yet, the tests do not tell us that the U.S. banking system is safe. The level of capital remains well below recent estimates of what would have been needed to avoid most of the financial crises in OECD countries over the past few decades. And the definition of capital (based on U.S. GAAP accounting) still fails to capture the risks of the largest banks with trillions of dollars of gross derivatives exposure.
Moreover, compared to the increasingly stringent U.S. regime, the European assessment is still in its infancy. So far, the Single Supervisory Mechanism (SSM) has published only one round of stress testing and—while far superior to earlier versions in Europe—it was still less stressful and less informative than those in the United States. The second round of tests—conducted jointly with the European Banking Authority—is due to be released early in the third quarter of 2016. However, the shocks to asset risk premia in the 2016 adverse scenario are significantly smaller than the euro area experienced in the financial crisis, limiting the usefulness of the results.
Importantly, even the best stress tests are no more than point-in-time assessments of specific shocks. There is no real-time transparency of bank net worth. Consequently, following the Brexit referendum, we don’t know which leveraged intermediaries, if any, suffered large losses and might now be severely short of capital. For example, we don’t know who sold insurance against the risk of heightened volatility in currency and capital markets. We don’t know who is exposed to an even longer period ahead of low interest rates, or who might face extreme losses from a U.K. recession, from a potential plunge in the value of London real estate, or from persistent economic weakness on the continent. And, we don’t know who might suffer losses from Brexit political knock-ons reflecting anti-globalization movements elsewhere in Europe and the United States.
What we do know—and this is useful information—is that European bank stocks (measured by the bank component of the Stoxx Europe 600) plunged by more than 20 percent (over two days) following the outcome of the referendum. More broadly, as of July 1, 2016, European bank stock prices were down by nearly one third since the start of the year and by more than 75 percent since peaking in 2007.
This suggests that capital in the European banking system is scarce; too scarce, in fact, to be confident about financial stability.
To get some sense of the magnitude of the capital shortfalls, we can look at the NYU Stern Volatility Lab’s estimates of systemic risk (SRISK). Rather than using accounting measures that lag big shocks like the Brexit referendum, the V-Lab relies on the market price of a bank’s equity to gauge a financial institution’s leverage. Specifically, for each listed intermediary, they compute the ratio of the sum of fixed liabilities and market equity to market equity. The logic is that equity investors are forward-looking, so real-time market valuations of bank stocks reflect perceived changes in balance sheet health.
More precisely, SRISK is an estimate of the aggregate shortfall of capital in the financial system in the event of a crisis (defined as a large—in this case, 40%—drop of the global equity market over the next six months). Analogous to stress tests, the idea behind SRISK is that an intermediary contributes to systemic risk to the extent that it is short of capital at the same time that a system-wide shortfall occurs (see, for example, here). It is precisely in such system-wide capital droughts that funding liquidity becomes most precarious, leading to damaging fire sales. To measure an intermediary’s capital shortfall, SRISK uses a benchmark: namely, 5.5% of the sum of equity and fixed-income liabilities in Europe, and 8% elsewhere. (This regional difference is a proxy for international accounting disparities, which tend to result in higher leverage values in Europe. See our recent discussion here.)
So, what does SRISK tell us about the potential systemic risk from Brexit? First, while down from the 2009 peak of more than $2 triilion, as of June 30 aggregate European SRISK surpassed $1½ trillion. That was its highest level since 2013, and was nearly three times higher than 10 years ago (see Figure 1). By comparison, while somewhat above the average of recent years, U.S. SRISK of less than $½ trillion was down by more than 50 percent from its 2008 peak. Second, European countries accounted for eight of the global top-10 list of financial systems ranked by the ratio of aggregate SRISK to GDP, with the United Kingdom in the number two spot, just behind Japan (see Figure 2 below). By this measure, the United States is not even in the top 20. (We scale by GDP to gauge both the importance of the capital shortfall in a country’s economy and the scale of resources that might be called on should the country decide to bail out banks in a crisis.)
Figure 1. Europe and United States: SRISK (Billions of U.S. dollars), 2006-June 30, 2016
Figure 2. SRISK by Country (Percent of GDP), July 1, 2016
The level and evolution of SRISK at specific institutions also is revealing. Table 1 lists the top 10 contributors to aggregate European SRISK as of July 1, 2016. Notably, four of the top five are based in France and Germany, a prevailing pattern in recent years that the Brexit referendum did not alter. At the head of the list, BNP Paribas has SRISK of $113 billion, representing 7.4% of the European total. The bank’s high SRISK reflects the combination of: (1) a large balance sheet (approximately €2 trillion); (2) high leverage (43); (3) high equity price volatility (112), and; (4) a high conditional correlation with the global stock market (0.76). At the same time, four of the top 10 contributors to European SRISK were U.K.-based banks, at least partly reflecting their elevated equity price volatility in the presence of Brexit uncertainty.
Table 1. Europe: Intermediaries Ranked by SRISK (Pct. of Total and U.S. dollars in billions), July 1, 2016.
Table 2 shows the top 10 European intermediaries ranked by the change of SRISK in billions of dollars over the period from June 17 to July 1 (with the referendum on June 23). Over this interval, the largest increase of SRISK—$15.6 billion—was associated with Lloyds Banking Group PLC. Lloyd’s SRISK rise can be decomposed into three parts: the increase of accounting debt during the period (Δ(DEBT), $0 billion), the drop in equity market valuation (Δ(EQUITY), $4.5 billion) and the rise in risk attributes of the firm (Δ(RISK), $11.1 billion). The latter reflects both the change in equity market volatility and the high conditional market correlation, with increased volatility the dominant factor in the Brexit episode. Notably, four of the top five institutions are based in the United Kingdom, while two of the top 10 are based in France, two in the Netherlands, and one each in Italy and Spain. Of the $84 billion aggregate rise in SRISK for these 10 institutions, roughly one fifth ($16 billion) directly reflects the plunge in their equity market valuation.
Table 2. European Intermediaries Ranked by Change in SRISK (Δ SRISK) from June 17 to July 1, 2016, Billions of U.S. dollars
To be sure, SRISK in Europe is below its crisis peaks. But taking comfort from higher capital levels and improved risk management since the financial crisis would be perilously complacent. Who in 2008 expected that AIG would have taken such concentrated risk (by insuring other intermediaries in the CDS market against lending losses) that it needed a Fed bailout to avoid failure? With more than $530 trillion (notional) of OTC derivatives outstanding at the end of 2015, it remains possible for a large intermediary to take on and conceal a dangerous level of systemic risk.
And there may be reasons to worry more today about financial fragility than in the past. Prior to the euro-area crisis that peaked in 2012, many European intermediaries (rightly) assumed that governments would bail out their financial counterparties. However, one consequence of the crisis has been the establishment of minimum requirements for “private sector involvement” (PSI) of creditors in the resolution of insolvent intermediaries. While these “bail-in” rules encourage more effective market discipline, they could also lead creditors to limit or halt credit supply more quickly. Put differently, under the new rules, the motivation for liability holders to run on fragile intermediaries and for financial institutions to dump risky assets ahead of others is now even greater.
What to conclude? Whenever big asset price changes raise questions about losses by leveraged intermediaries, there is a risk of a credit supply disruption, so investors and policymakers need to be prepared. It is true that capital in the financial system is substantially higher than it was in 2007, when it was virtually zero. Yet, as we have argued elsewhere, in our view current levels of capital remain insufficient to make the global financial system safe. This appears to be especially true in Europe, where stress testing and re-capitalization have seriously lagged U.S. efforts.
That said, we hope that Brexit turns out to be a non-event from a systemic perspective, and that authorities and bankers will implement the big changes (increasing capital and liquidity buffers) still needed to make Lehman moments no more than a distant memory.
Acknowledgements: We thank NYU Stern V-Lab’s Robert Engle and Robert Capellini for helping us to understand SRISK, and for providing the data used in this post. We also thank NYU Stern’s Larry White for helpful comments.