There are a couple of questions I have about imposing penalties on repo fails.

  1. Is the repo fail penalty imposed only on the seller of securities if it does not buy the securities back or also on the buyer if it does not return the collateral?
  2. I read that a lack of penalties sets the repo floor at 0%, I don't get why this is so.

In the US, a fail penalty is applied on the failure to deliver securities in a US Treasury or Agency transaction. This applies to any trade in these securities , including a simple cash trade , or the opening leg of a repo, or the closing leg of a repo.

The intent of the system , first introduced in 2012, was to clean up fails in the Marketplace by providing an incentive to settle. The convention in the marketplace is that if there is a settlement failure, then further daily attempts to settle are made until the trade clears. When this happens , the cash amount instructed doesn’t change over time. Therefore, a counterparty that fails to deliver securities is effectively making a decision to lend money at zero interest. But in the absence of a fails penalty, this may be logical if interest rates are close to zero and/or the securities are scarce.

In addition, In the absence of a fails penalty, repo cannot trade below zero. Why? There’s no point in lending money via repo at a negative rate in order to borrow securities that you have sold short. It is better economically just to fail on the cash trade.

  • $\begingroup$ If I understand correctly, In the case that there is fail penalty of 3%, in the opening leg of the transaction if the seller receives cash and fails to deliver securities it will have to give the cash back and pay 3%. In the closing leg if the buyer fails, it will keep the securities but will pay 3% meaning that the interest rate on that transaction was -3%. In the case of no penalty, the seller would just give the cash back and the buyer would just keep the securities which is as if they had made a cash transaction with 0% interest rate. $\endgroup$ Jul 11 '20 at 1:21
  • $\begingroup$ This also implies that the closer the repo rate is to 0%, or the scarcer the security the more likely is for the repo to fail. $\endgroup$ Jul 11 '20 at 1:22
  • $\begingroup$ This is a good answer, though the last paragraph isn’t exactly right. There’s no particular zero bound; the better way to think about repo pricing is relative to the GC rate. The GC rate can itself be negative (which happened occasionally before the Fed set up its RRP program). The best way to put is that specials tend not to trade below the GC rate less the fails charge, except under certain rare circumstances. $\endgroup$ Jul 11 '20 at 22:17
  • $\begingroup$ @dismalscience are you saying that the repo rate floor of 0% only works when the gc rate is positive? Or that there is never a repo rate floor? $\endgroup$ Jul 12 '20 at 5:55
  • 1
    $\begingroup$ @user3181821 Yes, that’s correct, and it’s what dm63 is pointing out. I’m just making the somewhat pedantic point that when the GC rate is negative and there’s no fails charge, the floor and the ceiling are basically the same (which is still below zero, though people will fail on trades constantly, which, as noted, starts to happen if there’s no fails charge and the rate is close to zero). $\endgroup$ Jul 12 '20 at 12:41

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.