Liquidity risk premia in unsecured interbank money markets

Liquidity risk premia in unsecured interbank money markets Jens Eisenschmidt, Jens Tapking

Main empirical observations 2

Main empirical observations Euribor: (unweighted) average of (up to) 43 individual rates, each rate reported by a socalled Euribor panel bank. The 43 panel banks are supposed to report "to the best of their knowledge [. . . ] rates being defined as the rates at which euro interbank term deposits are being offered within the EMU zone by one prime bank to another at 11. 00 a. m. Brussels time ('the best price between the best banks')" 3

Main empirical observations In „normal times“, spread of Euribor rate over repo rate should not significantly exceed spread of CDS on best bank(s). Otherwise arbitrage: Lend to best bank (rather than in repo market) and buy protection against a default of best bank. [Refinance in repo market if necessary. ] 4

Main empirical observations 5

Main empirical observations 6

Main empirical observation Is liquidity hoarding an explanation? ØUnsecured term money markets dried out with the start of the turbulence However: ØEonia rate still low (but volatile) and overnight market liquid: Banks do lend to each other unsecured, but only short-term ØRepo market rates still low and repo markets relatively liquid, including term repo markets: Banks do lend to each other long-term, but only against (good) collateral 7

Theory: basic idea Banks with a (cash) surplus will be reluctant to lend (unsecured) in the term market if Ø it is likely that they will face a liquidity shock (cash deficit) before the loan matures; Ø and can then only borrow at relatively bad conditions e. g. because (i) they do not have enough good collateral to borrow secured (ii) they have a high probability of own default. They will demand a high interest rate for term loans or prefer to lend overnight repeatedly until they indeed experience a liquidity shock. 8

Theory: assumptions 0 Ø B-banks have 2 -period deficit d 0 Ø L-banks have surplus d 0 Ø L-banks lend (unsecured) to B-banks for 1 or 2 periods Ø 2 -period CDS market (on B-banks) 1 Ø(1 -γ)(1/2) L-banks face shock +d 1 Ø(1 -γ)(1/2) L-banks face shock -d 1 - need to borrow - more if 2 -period loan in 0 - more expensive if little collateral or default risk 2 Ø Loans are repaid unless default Ø B-banks default with p. B Ø L-banks default with p. L ØB-banks borrow again if 1 -period loan in 0 9

Theory: results If Ø Ø Ø probability of liquidity shock in t=1 is zero (γ =1) or probability of default of L-banks is zero (p. L = 0) or L-banks have sufficient collateral then • spread between (2 -period) secured and unsecured rate (approximately) equals CDS spread (no funding liquidity risk premium); • there may be trading in unsecured term (i. e. 2 -period) market. 10

Theory: results Otherwise there is no trading in unsecured term (i. e. 2 -period) market, i. e. Bbanks only borrow (twice) for 1 period; 2 -period unsecured loans are offered at a spread over 2 -period repo rate that exceeds the CDS spread (funding liquidity risk premium); but B-banks would have demand for 2 -period unsecured loans only if there was no funding liquidity risk premium. 11

Final remarks Suppose that B-banks have enough collateral to borrow secured in t = 0. Then L-banks would be ready to lend for 2 periods because: re-use of collateral. Why disappeared funding liquidity risk premium recently? Because of ECB’s fixed rate full allotment policy (and expansion of eligible collateral) commitment? 12