How do you stop the worryingly high number of repo fails? Penalise the non-deliverers!
From Bloomberg (HT Yves Smith at Naked Capitalism):
Treasuries are in such high demand that investors are lending cash for next to nothing to obtain the securities as collateral through so-called repos, which dealers use to finance their holdings. The problem is many parties involved in repos aren’t delivering the bonds because there is no penalty for not doing so, causing “fails” to exceed $5 trillion, according to the Federal Reserve Bank of New York.
Now, an industry group is trying to fix the mess, which New York Fed Executive Vice President William Dudley said could cause U.S. borrowing rates to rise if not rectified. The Treasury Market Practices Group wants to impose a ‘penalty’ on failed trades, a move that may result in borrowers who put their Treasuries up as collateral for loans effectively receiving 2 percent interest.
As a reminder, repurchase agreements are one of the few remaining sources of liquidity in the financial system and a vital form of short-term financing for many market-makers. They’re essentially a sale of securities with an agreement to repurchase the same securities at a later time — sort of like a collateralised loan. A “repo fail” happens when a seller doesn’t deliver the promised security at the promised time. In practice they happen when it becomes cheaper to hold on to the securities than deliver them, or something like this, from a Federal Bank of New York paper:
Interest rates on special collateral RPs nearly always stay above zero because, instead of lending money at a negative interest rate to borrow a particularly scarce issue, a short seller can choose to fail on its delivery obligation… the cost of failing is about the same as the cost of borrowing a security on a special collateral RP at an interest rate of zero. It follows that failing is usually preferable to borrowing a security at a negative specials rate.
What’s interesting about the above Bloomberg story, is that the actions being considered could result in negative repo rates, that is, the lender not only lending money but actually paying the borrower to take that money.
This is something we’ve seen a few times before, most notably in — you guessed it — Japan. From Bloomberg again:
The Bank of Japan’s decision to adopt a zero interest-rate during the ‘Lost Decade’ of the 1990s because of deflation and a protracted banking crisis triggered the phenomenon for Japanese government debt. Rates less than zero surfaced in the U.S. in 2003, when the Fed’s target interest rate fell to 1 percent and traders sought to cover bets against 10-year Treasuries after their yields jumped more than a percentage point in about a month.
In the 2003 instance, dealers concluded it would be cheaper to pay interest to borrow the notes needed to fix their settlement fails than to continue to incur related-charges — not to mention annoying their customers. In Japan, the repos, which generally tend to trade close to the official interest rate, turned negative as Japan’s rates reached zero — suggesting little value or incentive to lend the securities.
As with breaking the buck in money market funds then, low interest rates are in danger of increasing the number of repo fails. The connection is also a good illustration of the Fed’s limits in easing interest rates. From Felix Salmon, who picked up the point late last month:
… below 50bp we’d be likely to get many more repo fails — which would significantly damage one of the few remaining sources of reliable liquidity. Which implies, given the way in which real-world interest rates have barely budged in the face of Fed easing, that the downside to cutting past 50bp is significantly greater than any upside. There might be another half-point in rate cuts in the offing, but beyond that, the Fed will probably cut no further.
Related links:
How low rates break the buck - FT Alphaville
Repurchase Agreements with Negative Interest Rates - Federal Reserve Bank of New York
______
