The “Greatest” Carry Trade Ever? Understanding Eurozone Bank Risks

Acharya, Viral V. and Sascha Steffen, “The “Greatest” Carry Trade Ever?  Understanding Eurozone Bank Risks,” The American Finance Association 75th Annual Meeting, Boston (2015).

Purpose:  To investigate the risks assumed by Eurozone banks as a form of carry trade, where banks loaded positively on GIPSI (Greece, Ireland, Portugal, Spain, and Italy) bonds and negatively on German government bonds.

Findings:

  • Banks’ stock returns were positively correlated with GIPSI bond returns, and negatively correlated with German bond returns, for the period January 2007 to June 2012.
  • Banks were effectively financing long-term GIPSI bond holdings with short-term German debt.
  • This carry trade behavior was more notable at large banks, banks with lower capital ratios, and banks with riskier asset portfolios.
  • Banks increased the magnitude of their carry trade between March and December 2010, ruling out the possibility that they were passively caught up in the crisis.
  • Banks’ carry trade exposure was related to actual bond holdings rather than to holdings of other asset classes.
  • Both GIPSI and non-GIPSI banks were involved, so the carry trade is not a case of GIPSI banks suffering from the weakness of their home countries.
  • Regulatory arbitrage was a likely contributor to the carry trade of both GIPSI and non-GIPSI banks.
    • Governments had incentives to maintain the Basel II zero risk-weighting on sovereign bonds so they could continue to borrow.
    • Banks with low capital ratios were incentivized to buy these zero risk-weighted sovereign bonds.
  • Risk-shifting by GIPSI banks may also have been a contributor to the carry trade of GIPSI banks.
    • An Italian bank, say, might have wanted to shift risk to a state of the world where they would be in trouble anyway (an Italian default) by buying Italian bonds.
  • Moral hazard may have also played a role, where banks in strong sovereigns might assume the risk of the carry trade with an implicit expectation of being bailed out in a worst-case scenario.
  • Moral suasion may have occurred where weak sovereigns convinced their home banks to buy own-sovereign debt.
  • U.S. money-market funds played an important role in providing or not providing liquidity to European banks.
    • After November 2010, they withdrew 60% of their investment in weakly capitalized banks and doubled their investments in well-capitalized banks.

Data:

  • Data on stock prices, bond yields, and CDS credit spreads comes from Bloomberg.
  • Banks’ portfolio holdings data is from the European Banking Authority (EBA).

Methods:

  • Regress each bank’s daily stock returns on the daily returns of 10-year GIPSI government bonds (all five countries), the daily return on 10-year German government bonds, and the daily return of bank’s home equity market (orthogonalized to the sovereign bond returns of Germany and of the home country).
    • \beta_{Greece} estimates each bank’s exposure to Greek sovereign debt, etc.
    • \beta_{Germany} is the estimate of each bank’s exposure to German bonds (a negative value indicates the bank is “short” German government debt).
  • Estimate the factor loadings quarterly.

Conclusions:

  • European regulators should not consider the sovereign debt of all Euro countries risk-free (or even high-liquidity).
  • ECB lending should probably be used to recapitalize banks rather than to guarantee their debt and encourage them to increase asset exposures.
    • While this props up the financial markets, especially in the banks’ home countries (due to home bias), it potentially makes a future sovereign debt crisis even more dangerous.
  • When banks hold long-term risky assets and rely on short-term funding, they are likely to encounter liquidity shortages.