The more different ways this argument is restated the better the chance is that readers will understand what is being said.

So, in what may be my last act of "advising", I'll advise you to cut the jargon. -- My old PhD advisor, to me, 26/2/11

the less healthy banks don't have excess reserves to deposit at the ECB and, in fact, have to borrow at the end of the maintenance period in order to have enough reserves to meet requirements. This drives up the rates in the heartbeat spike pattern shown.

The "unhealthy" banks probably have to borrow enough at the repo rate to cover their reserve requirements without resort to the interbank market at all.

One possible explanation of the heightened volatility and raised level in the past 8 months is that on the one hand the ECB is withdrawing extraordnary liquidity support (e.g., the 1-year tender wasn't renewed), and on the other hand every time an ECB official opens their mouth the market has a fit.

So, in what may be my last act of "advising", I'll advise you to cut the jargon. -- My old PhD advisor, to me, 26/2/11

If I understand correctly, the baseline rate of EONIA is currently established by the rate that the ECB PAYS depositors and that this is likely because the healthy banks are reluctant to lend overnight at higher rates to bad banks and instead deposit excess cash at the ECB for the minimal depository rate. THEY WILL LEND OVERNIGHT TO GOOD RISK BANKS, but, due to the relative surplus of cash compared to creditworthy borrowers, this tends to be very close to or at the depository rate. This establishes the  (green line) depository rate as the baseline for EONIA.

This is my reading too.

But the less healthy banks don't have excess reserves to deposit at the ECB and, in fact, have to borrow at the end of the maintenance period in order to have enough reserves to meet requirements. This drives up the rates in the heartbeat spike pattern shown.

This is not my reading, but it is not impossible that you are correct.

My reading is that EONIA is now completely decoupled from the MRO. Meaning that the bad banks don't borrow on EONIA at all - they borrow directly from the ECB. Which is why the ECB's liquidity support is being tapped so heavily. Which is what is getting Die Seriöse Leute's knickers in a twist. Under this reading, the spikes are due to banks being unable to predict precisely how much reserves they'll need. Since the banks who have excess reserves at the end of a maintenance period can always just leave them at the ECB for 0.25 %, they have the banks who underestimated their reserve needs over a barrel, and can demand more than 0.25 %

(Whereas in the normal course of operations, where EONIA is centred around the MRO rate, the banks needing reserves will sometimes have the banks with excess reserves over a barrel, because the banks as a whole overestimated their reserve needs, and there is room for EONIA to drop.)

But your reading is also possible: It is possible that the good (old boys') banks who are not generally prepared to lend overnight to the bad banks are willing to do so on the final day of the maintenance period, for some reason. Off the top of my head, I can't think of any, but that does not mean that one does not exist. The only way to really tell the two hypotheses apart would be to look a the volume data. But unfortunately that's proprietary.

If anything happens that decreases the reserves of most [good] banks, then fewer banks will have excess reserves to deposit at the depository rates and more will have to borrow reserves for longer periods of time. This increased demand for overnight loans will drive up EONIA and keep it up longer, depending on the severity of the problem.

Yes. With the caveat that draining liquidity should also thin the volume on EONIA, and thinner volume would be expected, all else being equal, to lead to increased volatility. And since EONIA is bounded at the low end by the deposit rate, the volatility would have to manifest in a trend towards higher rates.

- Jake

Friends come and go. Enemies accumulate.