Download
1 / 60
610 likes | 836 Vues
The NYU Stern Systemic Risk Rankings. Robert Engle, NYU Stern School of Business ECB Conference Macro Prudential Regulation as an Approach to Contain Systemic Risk September 2010. LESSONS FROM THE CRISIS.
E N D
The NYU Stern Systemic Risk Rankings Robert Engle, NYU Stern School of Business ECB Conference Macro Prudential Regulation as an Approach to Contain Systemic Risk September 2010
LESSONS FROM THE CRISIS The failure of large complex financial institutions can impose costs on the whole economy When they are failing, governments are in a compromised position. Unless there are liquidation or resolution mechanisms, governments need to rescue these firms. The potential of such a rescue reduces market discipline leading to excessive risk taking. Regulation of systemically risky firms is needed. But how can they be identified?
HOW CAN FIRMS CAUSE SUCH RISK? HOW CAN WE MEASURE THESE RISKS? COULD WE HAVE PREDICTED THE FIRMS THAT WERE THE CASUALTIES OF THE LAST CRISIS? WHICH FINANCIAL FIRMS POSE THE GREATEST RISK TO THE GENERAL ECONOMY?
STERN VIEW OF DODD-FRANK Forthcoming in October 2010
FUNDAMENTAL CAUSES OF THE FINANCIAL CRISIS • Risk was underestimated by many market participants (traders, money managers, bank ceo’s and boards, ratings agencies, regulators, investors and probably risk managers) • Many of these had strong incentives to ignore risks.
TWO KINDS OF RISK • INDIVIDUAL RISK • SYSTEMIC RISK
REGULATION • Regulate to reduce systemic risk, not all risk • Tax or capital charges on biggest, most systemically risky firms • Tax rate is countercyclical – higher when economy is doing well • Coordinate globally – potentially with teeth. WFO like the WTO, WHO? • Establish legal resolution authority to wind down complex international financial institutions in bankruptcy.
Market Structure of OTC (Over-The-Counter) Derivatives • “Too big to fail” is based partly on counterparty risk • Reduce counter party risk by moving popular OTC products to centralized clearing • The central counterparty will establish margin and collateral arrangements to protect itself from insolvent parties. • Central counterparty may itself have systemic risk. • Not all products can be moved. New and unique products will remain OTC. How can these important products be derisked?
OTC TRANSPARENCY – (Acharya and Engle) • Investors enter OTC contracts with assessment of risks including counterparty risk • However because of the opacity of this market, estimates of counterparty risk are noisy. • If counterparty exposures were public at say a weekly level, then OTC prices would include a counterparty risk premium. • Reduce counter party risk by trade transparency. • Centrally cleared products might be preferred by traders and dealers who do not want this level of transparency.
IT IS TIME • It is now time to put new regulatory structures in place. • It is time to coordinate this process globally. • As the finance sector recovers, there is a temptation to return to business as usual. • We cannot forsee the next crisis so we need robust institutions and appropriate incentives.
VOLATILITY INSTITUTE • I am the director of the Volatility Institute at NYU’s Stern School of Business • We are the host of SoFiE which just had its third annual meeting in Melbourne and next year in Chicago. • We have research associates all over the globe. • We host the Volatility Laboratory or VLAB
VLAB.STERN.NYU.EDU Over a year ago the Volatility Institute introduced the Volatility Laboratory – an online web site with daily estimates of volatilities and correlations for a wide range of assets and methods. It has now grown to include hundreds of assets in: Commodities Equity Sectors Equity Names International Equities Exchange Rates Treasuries Corporates Real Estate Volatility Indices
RISK MEASURES • Calculation of volatilities and correlations on a daily basis allows updated risk estimates for firms, sectors, asset classes and countries.
INTRODUCING THE RISK PAGE Now we introduce a page providing estimates of risk for the 102 largest US Financial firms. Risk is estimated both for the firm itself and for its contribution to risk in the system. This is called the NYU Stern Systemic Risk Ranking. This is updated weekly/daily to allow regulators, practitioners and academics to see early warnings of system risks.
SYSTEMIC RISK “Financial institutions are systemically important if the failure of the firm to meet its obligations to creditors and customers would have significant adverse consequences for the financial system and the broader economy.” Daniel Tarullo Federal Reserve Governor
MEASURING SYSTEMIC RISK Acharya, Pedersen, Philippon, and Richardson(2010) propose the use of market data to estimate systemic risk contributions of firms. Their central measure is Marginal Expected Shortfall or MES. This paper develops a new dynamic methodology to estimate MES from equity data. MES is combined with leverage and size data to measure systemic risk contributions.
LEVERAGE • Highly levered firms have a greater risk of default. • The default of a firm is far more dangerous if the economy is weak and highly levered as there are no buyers to assume the liabilities. • Thus firms that are considered systemically risky are firms that face capital shortages just when the financial sector as a whole is capital constrained.
“LEVERAGE EXTERNALITY” • High leverage is only dangerous for the economy when everyone is doing it – this is an externality! • This is why regulation is required.
THE APPROACH We want to estimate for firm i, the loss in a future crisis: As we have little data on crises, it is necessary to carefully structure the problem. Estimate the expected equity losses for a firm from a modest decline in overall returns. Estimate equity loss and capital shortfall for leveraged firms in a crisis. NYU Stern Systemic Risk Rankings are the percent contributions to total capital shortfall.
Volatility, Correlation and Tails for Systemic Risk Measurement Christian Brownlees and Robert Engle Stern School of Business
MARGINAL EXPECTED SHORTFALL The expected shortfall of a market index is defined by ES is a useful and coherent measure of risk. Recognizing that the market return is a weighted average of individual firm returns, MES can be interpreted as each firm’s contribution to system losses.
INTERPRETING MES • In words, MES is the expected loss incurred by equity investors in a firm, when the general market suffers a big decline. • We will often use a 2% daily market decline to measure MES. Hence the market expected shortfall is a number greater than or equal to 2. It is higher when volatility is high. • Firms with MES much bigger than 2, are the biggest losers in a market downturn.
Use flexible time series approaches to modeling volatilities, correlations and tails. The Model: Disturbances are serially independent, mean zero, variance one, uncorrelated but not independent random variables. Copula. Volatilities are Asymmetric GARCH models Correlations are Asymmetric DCC.
THE CALCULATION At time t, MES is given by Firms are risky if they have high volatility Firms are systemically risky if they also have high correlations. Market ES is the same for all firms Estimate tail probabilities non-parametrically
DATA • Unbalanced panel of 102 large U.S. financial firms 1990-2008 • Firms in 4 industry groups • Depository Institutions • Insurance • Security and Commodity Brokers • Others • Market Index • Quarterly Data from Compustat on Quasi-Leverage.
CROSS SECTIONAL FORECASTING On a day when Rm<C, what is the rank correlation between Eq lossi,t and MESi,t? How accurate is the cross sectional distribution of losses. Construct a Gini coefficient between MES and future losses.
Equity Loss in Crisis • To estimate the fall in equity value in a crisis, an adjustment is made to MES • MES is adjusted to measure the expected fall in equity prices that would occur in six months if the market return is worse than a 40% decline. • Approximately this is 18 times daily MES.
MULTI-STEP FORECASTING • Simulate the bivariate outcome of (ri,rm) for six months starting on date t using the estimated model for volatilities, correlations and copula. • Examine all the scenarios where market return falls by at least 40%. Find average loss for firm i. • Average loss in a six month crisis/average loss in a 2% down day is ~~18. More precision will come later.
SRISK% : NYU STERN SYSTEMIC RISK RANKING As equity values fall in a crisis, leverage increases until the firm is in distress. Nominal debt is taken from compustat and changes little over time. DISTRESS=min(0,Eq-kA)=min(0,(1-k)Eq-kD) where k is a prudential standard ratio of equity to assets such as 8%. Eq is equity in a crisis, D is debt and A is asset value = Eq+D
SRISK% When equity falls dramatically, distress becomes positive and it is the contribution by each firm to aggregate distress that is the Stern Systemic Risk measure. SRISK%i,t= Distressi,t /(Total Distresst )
