Friday, August 10, 2012

Pushback on the IOER Debate

Should the Fed lower the interest on excess reserves (IOER) to help stimulate the economy?  Cardiff Garcia of FT Alphaville and I recently discussed this question.  He made the case that lowering the IOER is likely to severely disrupt short-term financial intermediation, particularly money market funds (MMFs), and its benefits are not clear.  I countered that with the right signalling the Fed could meaningful add monetary stimulus by lowering IOER without disrupting MMFs.  Garcia responded by questioning whether the Fed could really deliver what I claimed and to this point I replied here.

Now it is Dan Carrol's turn.  He says not so fast Garcia and provides counterarguments.  Let me add to Carrol's points that the Treasury is now gearing up to handle negative interest rate bidding.  Bloomberg reports the following:
The Treasury also said it is “in the process of building the operational capabilities to allow for negative-rate bidding in Treasury bill auctions, should we make the determination to allow such bidding in the future.”
Now go read Carrol's critique of Garcia's assessment. I have also posted his remarks below the fold.
Generally, [Garcia's] articles focus on the screwy things that happen when interest rates hit zero. He starts out well, but then gets highly speculative. The key issue here is this: is preserving an artificial rate that is north of zero to protect the functioning of money market funds the right monetary policy?
Where he goes off track is his speculation that negative rates would result in deflationary expectations. It is well understood that the causation runs in the other direction. Otherwise, we could stimulate the economy by raising rates.
His argument rests on the premise that collapsing the Fed Funds rate to zero might result in a panic and a significant disruption of the financial markets (specifically money market funds), might result in negative interest rates and a reduction in supply for money market investors, and might result in a disruption of the Fed’s t-bill auctions. Let’s deal with those in turn:
  1. Money market funds charge fees and need a yield to cover their fees, so that they can continue to pretend that their services are “free” to investors. ZIRP essentially exposes a flaw in their business model. I tend to think that problem is overplayed, other than paying lawyers to rewrite prospectuses to allow funds to charge fees in other ways, the way banks do (or allowing banks to gain market share at the expense of MMF’s). It’s not the Fed’s job to protect the business models of banks and money market funds, other than perhaps provide enough breathing room for a transition.
    1. In the extreme, banks might hoard physical cash. I’m not sure this argument is worth considering, as the storage and logistics costs are likely to be high. If this becomes a problem, well, then we have bigger problems than some MMF’s going out of business.
  2. If banks move their excess reserves away from the Fed and instead move to T-bills and other safe assets, the argument goes, that will drive yields negative as investor demand will bid up prices. The key flaw in this argument is that it is assuming that the move of eliminating IOER will be unsterilized. The Fed invests the excess reserves into T-bills and other Treasuries. If the excess reserves go away, one would hope that the Fed would sell off assets in response. Indeed, the Fed could simply give banks T-bills when they ask for their reserves, rather than pay a dealer to conduct a transaction.
    1. I am assuming that the Fed wants only to incrementally expand the money supply, rather than take an amount equal to all of the excess reserves and dump them onto the economy.
    2. The corollary to this argument is that the reduction in supply to money market funds could cause some funds to “break the buck”, as they could no longer buy assets and be guaranteed the return of principal. For a response, see both #1 and #2 above.
  3. The T-bill problem stems from the fact that the Treasury is reluctant to sell T-bill’s at a negative rate. That sounds like a personal problem to me. The US Government is borrowing at such a heavy rate that if they don’t sell T-bills, a lot of retirees are going to be disappointed. Most of the rest of the argument on the role of speculators is, well, speculation.
However, if the Fed is successful in raising the inflation rate, then interest rates will rise, and the point will be moot.
One objection to eliminating IOER is similar to one I read recently at Self-Evident. Normally a very astute observer of fixed income markets, she starts with the premise that “the theory behind reducing IOER is that it would encourage bank lending ... In reality, this will not be a useful tool if loan demand remains low …
The theory behind the reduction of the interest rate on excess reserves is not that it would encourage bank lending per se. It is that it would discourage the accumulation of excess reserves held at the Fed. From a monetary policy perspective, it matters not whether bank reserves are held in T-bills or lent out to businesses (it does matter, but that is a more complicated discussion). It matters greatly, however, if bank reserves are held at the Fed, which is akin to stuffing the money under a mattress, assuming the Fed doesn’t reinvest the money in T-bills. Money held at the Fed but not reinvested has a velocity of zero and contributes not to GDP. Even if the Fed invests the money, do we really want the Fed allocating capital for the economy?
The Fed has tossed out the idea of reducing IOER. As I see it, there are two lines of thinking from the Fed:

1.      The institution of IOER in the fourth quarter of 2008 is frequently blamed for deepening the recession and further disrupting the already disrupted short-term credit markets (by incentivizing banks to withdraw money from the money markets and park cash at the Fed, resulting in an explosion of the Fed’s balance sheet). So by offering it as a possibility, the Fed is subtly admitting to the poor timing of the policy while offering a way to make the persistent criticism go away.

2.      The huge balance sheet at the Fed is a political football. By incentivizing the banks to move excess reserves elsewhere, the Fed gets to shrink its balance sheet, a move that is sterilized if the Fed sells its holdings, while covertly engaging in other stimulative moves (such as selling fewer holdings than it should).
While there are legal constraints on what the Fed could do, the primary purpose of monetary policy is to increase the money supply. In theory, there are an unlimited number of assets that the Fed could buy to pump cash into the economy. Most directly, there are the foreign exchange markets. Unfortunately, this nuclear option is not popular with our trading partners, despite the fact that it is not really that different economically from buying Treasuries. Legal constraints could be changed if the political will existed to do so.
FT Alphaville discusses the arbitrage that is taking place, as banks borrow money from the GSE’s and then invest the money at the Fed, since the GSE’s don’t have access to the Fed’s IOER and therefore charge a lower rate. The banks then pocket the spread. If the Fed cut IOER, these arbitrage opportunities would be eliminated. This is cited as a problem because trading in the fed funds market could dry up, and therefore could perversely raise the effective fed funds rate.
I’m not sure why this is relevant, except that eliminating arbitrage opportunities for the banks should be a good thing. The goal of cutting IOER is to provide a disincentive for banks to park reserves at the Fed. The arbitrage results from an artificially high fed funds rate, which is a monetary tightening mechanism and a way for the banks to pick the pockets of the GSE’s. There is no reason the banks should receive a rate that is materially different on a risk-adjusted basis than the GSE’s, or anyone else for that matter.
As far as driving up the Fed Funds rate on a decline in volume, that simply means that the Fed Funds rate will cease to be a relevant measure of monetary conditions. Perhaps the Fed will then focus on something more relevant. I’m not convinced, however, that a decline in volume will necessarily result in an increase in the Fed Funds rate. The rate is contingent upon the balance between supply and demand, and more importantly the availability of alternatives. The effective fed funds rate would move in sync with other markets, though on low volume it might exhibit more volatility.

8 comments:

  1. I do not understand why the FT Alphaville folks (who are terrific at their jobs) are so apoplectic about disruptions in the MMF industry. Even if all the envisioned disruptions in eliminating IOER come to pass, why is this so catastrophic? The MMF industry is broken. We need to fix it. Blowing it up is one way to do that. The parking of excess reserves at the Fed is hurting the economy so it needs to stop. We need to continue on this path forever because we so afraid of a shock? How bad does it need to get before we summon the courage to try something new? Ten years of persistently high unemployment? Depression level unemployment in Europe continuing indefinitely? All because we are afraid of a MMF financial shock which may never occur and which, if it does, may be a good thing? People who are against eliminating the IOER are essentially saying the status quo is acceptable. It isn't. Take a risk, stop acting so afraid.

    ReplyDelete
  2. No matter what interest rate the Fed pays on reserves, the amount placed with the Fed (or kept on the bank's own books) won't change. That amount is simply a function of the Fed's balance sheet.

    The banking system as a whole cannot meaningfully reduce excess reserves. One bank's attempt at reducing the excess reserves it holds pushes it to another bank. This process continues until banks can longer find assets yielding more than the Fed pays.

    Banks are not constrained by reserves. Reserve management at banks is done by arbitraging short term (mostly overnight), low risk instruments like tbills, general collateral, etc.

    ReplyDelete
  3. I'm flattered that you posted this. Thanks.

    Ed,
    I believe the function you are describing is related to net reserves, not gross reserves (the Fed's assets and liabilities netted out).

    ReplyDelete
  4. Hi David, thanks for keeping the conversation going. This is a more difficult topic than many realize, and certainly reasonable minds will disagree.

    I'll respond to Dan in full over at FTAV soon. But I'll make two quick points here.

    One is that I've been very careful to note the uncertainty of what *could* happen if IOER were to be eliminated rather than claiming to know the definite consequences. There are real risks here that I think a lot of people are unaware of, and that should be included in any probabilistic assessment of the idea's benefits and costs. That's all.

    Second, I think Dan misrepresents about my point on rates. All I'm saying is that in a safe asset-deprived world, *negative* nominal rates might have a perverse and (counterintuitively) deflationary effect. I'd never make the nonsensical point that raising rates equals stimulative policy. Again there is a lot of uncertainty about negative rates and I'll write more later. I don't think Dan is deliberately trying to make me look like an idiot, but there are some nuances here about which we disagree.

    Anonymous, cheers for the note and the kind words (and please keep reading us!), but I assure you that we aren't apoplectic about anything. I'm happy to admit that MMFs and money markets generally need reforming and have written as much. But implementing the appopriate reforms needs to be done carefully, and in the meantime we have to deal with the money markets we have, not the money markets we hope to have one day. And the money markets we have are still systemically important and dangerous.

    Everyone, I'm genuinely not trying to harsh the market monetarist mellow here, and I agree with David on quite a few things.

    It's just that on the particular issue of IOER, I think that many economists (David is a noteworthy exception) haven't looked closely enough at the potential consequences, so we at Alphaville have tried to bring them to light.

    Anyways, keep the back and forth coming. This kind of iterative conversation is the essence of the blogosphere. Cheers, Cardiff

    ReplyDelete
  5. The banks won't continue to hold excess reserve balances. They will buy close substitutes (the MMMF's governments). I.e., just turn back the clock to the time when savers made more indirect investments (as opposed to direct investments).

    ReplyDelete
  6. Dan,

    I am talking about reserves plus excess reserves so I am talking total reserves. Banks in aggregate can do little to change how much they hold.

    Flow5,

    One bank can make that choice, but banks in aggregate won't be able to do anything about it.

    ReplyDelete
  7. Thanks Cardiff for the comments. Yes, this is a complicated matter with a lot of moving parts. My view after our previous conversation is that we should not do IOER change by itself. I am afraid for the reasons Cardiff lines out that by itself a IOER change may be too incremental and thus costly. It would need to be part of a bigger shock and awe package, like introducing open-ended QE tied to a NGDP level target.

    Dan Carroll, I see that you are from Austin. I just spent the past 5 years at Texas State University down the road from you. Love and miss the Hill Country already.

    ReplyDelete
  8. I'm guessing I was unable to prove I am not a robot ...

    (I've had trouble posting my comments)

    Ed, I misunderstood your comment, as I think you were making the same argument the NY Fed was making. The banks in aggregate can't reduce total reserves held at the Fed, but they can reduce excess reserves. In doing so, the velocity increases. Scott Sumner responded eloquantly to the NY Fed, and I posted a response as well.

    Cardiff, I was not trying to misrepresent you. As a finance practioner, I am accustomed to a style of debate where I sometimes carry an opposing argument to logical extremes. Often when we are done debating, we are forced to apologize. It is necessary break down egos because we often have money on the line. I'm not a big fan of negative rates, either, except in a very modest indirect sense, so I apologize if I got carried away. I was not just responding to you, though, but also to others who have repeated the same arguments.

    David,
    I had noticed that you had moved. Too bad, though Texas in August is a bear.

    ReplyDelete