Exit strategy, Part One: ZIRP
The Fed has announced plans to raise rates in the imminent future, but the market does not believe it. Why not? Conventional wisdom appears to be that the Fed will chicken out, just as it did during the so-called Taper Tantrum. The Fed has signaled its appreciation that “liftoff” will involve increased volatility, and has stated its resolve this time simply to let that volatility happen, but markets don’t believe it.
I want to suggest a slightly different source of disconnect, concerning expectations about what exactly will happen in the monetary plumbing when the Fed raises rates. Case in point is the recent Credit Suisse memo, apparently the first of a series, that forecasts “a much larger RRP facility–think north of a trillion” whereas the FOMC itself “expects that it will be appropriate to reduce the capacity of the [RRP] facility soon after it commences policy firming”. That’s a pretty big disconnect.
Pozsar and Sweeney (authors of the CS memo) think about the exit from ZIRP (Zero Interest Rate Policy) from the perspective of wholesale money demand, which they insist is “a structural feature of the system” and “the dominant source of funding in the US money market”. Before the crisis, that money demand was funding the shadow banking system, largely through the intermediation of repo dealer balance sheets. Now, it is funding the Fed’s balance sheet, largely through the intermediation of prime money funds and US bank balance sheets, both of which issue money-like liabilities and invest the proceeds in excess reserves held at the Fed.
The big problem that now looms is that neither prime money funds nor banks want that business any more. Capital regulations have made the bank side of the business unprofitable, and looming requirements that prime money funds mark to market (so-called floating NAV rather than constant NAV) will force them out of the business as well. Where is that money demand going to go? Pozsar and Sweeney say it will go directly to the Fed, causing the swelling of the Reverse Repo Facility pari passu with the shrinking of excess reserves. The mechanism will be a shift from prime money funds and bank deposits into government-only money funds, which will absorb the flow by accumulating RRP.
In other words, the Fed will not be able to shrink its balance sheet as part of this first stage of exit from quantitative easing. It will only be able to shift the way that balance sheet is funded–much less excess reserves held by banks, much more RRP held by government-only money funds. Nevertheless, because this shift will allow the Fed to regain control over the Fed Funds rate, it will accept that consequence. Exit from ZIRP comes before exit from QE.
Are you with me so far? If not, I urge you to have a look at Exhibits 1, 3, and 6 in the P&S memo. All I have done is to translate them into English.
I have some quibbles with details of the memo, but they do not concern the central point up to this point. One difference however is important for what comes next. Unlike P&S, I would not describe the shift from pre-crisis to post-crisis as involving shrinking the shadow banking system and growing the traditional banking system, as they do on p. 3. I would rather point out that the Fed itself took a sizeable chunk of the shadow banking system onto its own balance sheet. Reserve liability funding of RMBS lending=money market funding of capital market lending, which is my definition of shadow banking, as also theirs. Why does this matter?
Suppose everything happens just as they anticipate. What then? What then is that there will be two policy rates in the system, not one. The Fed Funds rate is the rate for the traditional banking system, and the RRP rate is the rate for the market-based credit system, at least for that fraction of the market-based credit system that sits on the balance sheet of the Fed. From this standpoint, what would eventual exit look like? Eventual exit would involve the Fed shedding its own shadow bank-like exposures.
How can that happen? Maybe first the Fed steps back from being money dealer of first resort and serves only as money dealer of last resort. In practical terms that means that it makes the outside spread, not the inside spread. And that means that it must offer an RRP rate that is below private wholesale money rates, indeed far enough below that wholesale money demand will shift over to funding private issue. Only then can the Fed take the second step, terming out its own funding or, more reasonably, packaging its long term assets into a special facility that is funded separately. Only then could RRP shrink, and with it the size of the entire balance sheet.
This seems to me a plausible road to exit (with one pretty big but not insurmountable obstacle, namely capital losses on the assets). Is that maybe what the Fed has in mind?
This way of thinking may shed light also on the disconnect between the Fed interest forecast and the market interest forecast that we mentioned at the top. Possibly that difference is not so much about market expectations that the Fed will chicken out, and more about an expected increasing divergence between IOER and other money market rates. P&S Exhibit 4 shows plainly that the Fed’s RRP rate has provided a floor while IOER has provided a ceiling; Fed Funds effective has been in the middle. The market thinks that IOER can go up with no consequences for other money market rates. Even worse, the market thinks that there are trillions of money funding about to be released from bank deposits and prime money funds, which if anything will be driving wholesale money rates down, not up.
For me, the money quote from P&S is the following, from p.2:
“The untold story behind the changes in the money market since 2008 is policymakers’ struggle to find a permanent home for the money demand of cash pools in a financially stable manner that is also consistent with control over short-term interest rates.”