Friday, December 16, 2016

October Inflation and the rate hike

Inflation continues its recent path.  Core CPI is at 2.1%.  Shelter inflation is up to 3.6%.  Core CPI inflation excluding shelter is down to 1.1% and the trend is down.  Core CPI inflation excluding shelter over the past 9 months has only risen by 0.4%.

The Fed, as expected, raised rates this week.  Tim Duy notes that the rate hike came with a bit of a hawkish shift.

At this point, I am cautiously maintaining my expectation that the most likely outcome here is a decline in long term interest rates and some economic contraction, but my certainty is not high, and this could play out over months.  If long term rates continue to rise sharply over the next month, then I will be less confident of that outcome.

Here is a graph of the Eurodollar yield curve.  Rates had actually started to rise somewhat before the election.  Then, there was a jump right after the election.  Since then, the curve has risen somewhat.  With the rate hike, the short end of the curve moved up, but the long end didn't move much.

With interest on reserves as the new monetary tool, and with the Fed balance sheet more or less stationary, a reaction from the banks of sending more reserves to the Fed would have to come at the expense of currency, if I am thinking about this correctly, so shrinking currency would be a bearish signal, I think.

This new monetary regime seems strange to me.  Normally, with the Fed Funds rate target, the Fed would basically be in control of the quantity of currency and reserves, and they could make daily adjustments to purchases around the target.  But, it seems like now they have created a situation where the quantity of reserves and currency will be a market function.  If they aren't planning on buying and selling treasury bills like they used to, it seems like quantities could shift quite a bit while we wait on the next FOMC meeting.  But, I'm not an expert on the nuts and bolts, here, so please tell me if I am incorrect.  Maybe it would take more than a few basis points for things to get out of whack, but do we know how elastic the supply of excess reserves is?


  1. I think those are great questions. The Fed is in uncharted territory. It never should have raised the rate of IOR (it shouldn't have ever paid IOR). It should have started to let its balance sheet shrink with the goal of getting back to the pre-2008 regime of using OMO to move the money supply and the Fed funds rate. I don't understand why it has chosen to keep the large balance sheet. Can you see any advantage to this change?

    1. The Fed's stated reason for paying IOR is:
      The Fed’s new authority gave policymakers another tool to use during the financial crisis. Paying interest on reserves allowed the Fed to increase the level of reserves and still maintain control of the federal funds rate. As the Board’s website states, “Paying interest on excess balances should help to establish a lower bound on the federal funds rate.”

      ...which sort of leaves me speechless.

  2. Egads--core CPI minus shelter is very low and falling! And the Fed, oblivious to zoning and housing markets, is tightening.

    The dollar soaring. Animal spirits higher on Trump but the U.S. monetary noose tightening. Already bad for Hong Kong property.

    2017 a cipher. Maybe we slog along again.

  3. Egads--core CPI minus shelter is very low and falling! And the Fed, oblivious to zoning and housing markets, is tightening.

    The dollar soaring. Animal spirits higher on Trump but the U.S. monetary noose tightening. Already bad for Hong Kong property.

    2017 a cipher. Maybe we slog along again.

  4. Hmm let me see if I am thinking about this right.

    With traditional OMO, the Fed regulates spending (NGDP) by regulating the quantity of base money in the economy. By selling a bond, the Fed is paying the investor to remove their money from the economy. Bond prices determine the amount the Fed has to pay to keep D dollars from being spent for X amount of time. Of course the investor can sell the bond sooner, but that's just moving base money around; if the market needs more base money sooner, tough. The Fed has to decide to re-create it (buying the bond back).

    With IOR, the Fed regulates spending (NGDP) by paying the market per day to keep base money parked in reserves. Here, the quantity of base money is fixed at a very large quantity, but the Fed pays people (through the banks) if they don't spend it. If you pay people enough, the money won't be spent, and NGDP is effectively controlled. If the investor needs the money sooner, no problem; they remove it from reserves and stop being paid IOR. But it's up to the investor, not the Fed, to re-introduce the dollar into the economy.

    It looks like the same effect (sidelining dollars to limit NGDP growth) for the same cost (the IOR rate or the short-term Treasury yield), with one big difference: who gets to decide when to re-introduce that dollar into the economy. This difference might be really important in a liquidity crisis. Under an OMO regime, the markets have to count on the Fed to create enough liquidity fast enough in the next financial crisis. Under an IOR regime, the markets know the cash is there if needed, so won't panic to the level we saw in 2007-2008. By giving the market more control over the size of the effective (non-sterilized) monetary base, the Fed is delegating some control. If the monetary base is leading to excessive NGDP growth, the Fed simply sterilizes it more with IOR.

    The more I'm thinking about this, the more I'm wondering if it isn't a fundamental improvement, to let the market "create" (de-sterilize) almost as much money as they want as fast as they want, using IOR to set the price of private money creation. The view here is that a dollar sterilized by IOR is not really part of the "effective" monetary base, so the market is deciding on the size of the "effective" monetary base, with the Fed guiding that decision by setting the IOR rate. In an IOR regime, the total monetary base is never too large. The Fed can create as many trillions of dollars as it wants, and then sterilize them with IOR all day long. As long as the IOR rate is less than inflation, an enormous-sterilized-monetary-base policy is effectively costless --- really cheap liquidity crisis insurance.

    Maybe something really great happened here that we haven't understood yet. Letting markets regulate the size of the effective monetary base might be really great.


    P.S. almost as great as free banking! (I'm kidding, I don't even understand free banking.)

    P.P.S. I'm not sure what happens if forecasts of excessive NGDP growth drive IOR above inflation. The Fed's IOR costs start to look pretty bad. Of course this is the same case where bond prices fall and OMO's start to become money losing. I guess the answer in both cases is to reduce the monetary base through taxation. Am I discovering the FTPL?

    1. Trying to think this through better, the part I realize I'm not understanding is how raising IOR sterilizes any dollars. They can still get spent; they just move around more. I guess my theory would have to be that higher IOR makes banks want to hold onto reserves and so less inclined to make loans, so we're back to the money multiplier as the channel?

    2. I don't know. I think the Fed is still basically controlling currency growth, but they are just adding a layer of uncertainty. With the Fed Funds Rate target, at least they could miss their rate target on a daily basis if they needed to diverge from it to meet the demand for money. But, IOR is a contractual term, so a change in the demand for reserves will lead to a change in reserves, and after setting the rate the Fed won't have control in the short term. Maybe that doesn't make much difference, or maybe they have short term mechanisms in place that I don't understand.

      But, it seems like in 2008 they set the FFR too high, and they were faced with a choice of either hitting their rate target or meeting the demand for liquidity. They would have had to sell a bunch of treasuries to keep FFR at 2%. For some reason, they felt that meeting their rate target was more important than meeting the demand for liquidity. IOR allowed them to take liquidity out of the credit channel instead of taking it out of currency, while keeping rates at 2%. Of course, rates were at 0.25% weeks later, so I don't understand why they aren't more widely criticized for their thinking on IOR in Sept. 2008. If the point was to keep rates at 2%, didn't it fail even on their own terms?

      It seems dangerous to me because, as with Sept. 2008, if the Fed sets the FFR too high, if they can't hit the target without sucking a bunch of cash out of the economy, it should be immediately obvious to them that the rate target is off. I guess that wasn't obvious to them in Sept. 2008. But, it seems like IOR just automates that lack of a response. It seems like if the banks are induced to want more reserves they will pull currency out of the economy without having to sell any treasuries to the Fed.
      But, I'm not 100% confident about my understanding here, either.

    3. Oops. That last part is a typo. The Fed would be selling treasuries, not the banks.

    4. > after setting the rate the Fed won't have control in the short term.

      But that's my whole point! With IOR, the physical MB can be much larger than the effective MB, because IOR sterilizes some reserves. The quantity of reservers that are sterilized varies based on market conditions. That means if a set of credit-worth institutions (say, institutions impacted by a sudden collapse in liquidity of MBS) suddenly need to borrow a truly large amount of money, then the banks can make those loans based on their assessment of the credit worthiness and long-term soundness of the borrower, without being constrained by reserves. We no longer need to rely on the Fed to swoop in and rescue, because the banks can do it themselves.

      > If the point was to keep rates at 2%, didn't it fail even on their
      > own terms?

      Policy was terrible in 2008 but consider two counterfactual policies:

      1) same as actual policy, but with IOR of 1% instead of 2% between May and December of 2008 (or whatever the exact dates were)

      2) same as actual policy, but with IOR of 3% instead of 2%, oh, and a statement that the Fed is repealing inflation targeting, and all future monetary policy will be driven by an NGDP level-path target of 2007 NGDP marching forward at 5% a year no matter what

      Which would have been more successful? Of course you know what I think and that I expect you to think the same. I belive policy was a disaster not because of the precise setting of IOR (0%, 1%, 2%, 3%, whatever), but because *no* monetary base expansion could have supported NGDP growth as long as the markets believed that the Fed would prefer an NGDP plunge over a small risk of above-target inflation. Think about how much more criticism the Fed would have gotten for allowing a few years of 3-4% inflation than it got for the worst NGDP performance since the great depression. The Fed got plenty of criticism as it was, and, if you read the WSJ, all of that criticism was for inflating a housing bubble and then taking on too much *risk* of inflation, when, as you know, its actual policy error was in allowing NGDP to plunge. It's quite amazing. Given that climate, the markets assumed correctly that the Fed would do whatever it took to keep inflation under its ceiling, including taking away the larger MB or paying enough IOR to sterilize it completely, and so the markets allowed NGDP to fall.

      Agreed that nothing I'm saying about IOR changes any of that. My point is that you're criticizing Fed's specific policy 2008 when I'm trying to focus on the improvements we can hope for in a new world where Fed policy is to keep a very large monetary base and then sterilize it with IOR. The improvement I'm hoping for is much better ability for the market to deal with financial crisis by itself, and not rely on Fed intervention like we did in 2008. I wasn't trying to claim that IOR has some magic ability to get the effective MB to be the right size to hit an NGDP level-path target. To do that, we need a Fed commitment to an NGDP level-path target as the overriding policy goal.

      > But, it seems like IOR just automates that lack of a response.

      Sure. IOR does not automatically adjust the MB to the most socially desirable level. But it does seem like a very big improvement to allow the banks to respond to Lehman and Bear Sterns themselves. As a commercial bank with a cool $T in excess reservers earning pennies in IOR, an opportunity to lend a lot at 2% or 3% to enable a fundamentally solvent institution like Lehman to survive looks pretty appealing. The Fed can delegate the rescue missions to commercial banks, which seems like a huge improvement compared to the 2008 world where only the Fed had the power to conjure needed reservers into existence.

    5. I think that's true, as far as it goes. But, reserves have not been the constraint. Capital has. Even back in 2007, there was famously a rabid appetite for AAA securities. There was plenty of demand for near cash securities. But, even with all that cash, the intrinsic value of those securities was below par because of expectations.

      That was the problem in late 2008, too. The Fed Funds rate was below the target rate because the banks were flush with cash and reserves. I don't think the Fed even has that as a stated reason for the policy. And the problem with their policy was that it did nothing to solve the capital problem. And, the capital problem was a product of expectations and collapsing NGDP growth.

      So, I think you're right that there wouldn't be a liquidity problem, but the reason nobody would lend to Lehman was because they considered them to be insolvent. That did mean that Lehman became illiquid, but putting a bunch of cash in the other banks' balance sheets wasn't going to solve that.

      It might have been possible for Lehman to be solvent in Sept. 2008, but that would have to come from NGDP and asset price expectations.

    6. Thanks Kevin. I guess I'm off in the weeds and over my head at the same time. I'm not sure what it means for "capital" to be a constraint. I know what it means to be liquidity constrained (no cash on hand) or balance-sheet constrained (everything of value you own is fully leveraged). Is capital constrained the same thing as balance-sheet constrained?

      Whether Lehman was balance-sheet constrained depends on how you value its holdings of MBS, which I guess is your point. I think you're right that plunging NGDP expectations were a big driver of falling MBS prices, and there's no reason to think larger reserves in March 2008 would have helped much with NGDP expectations. So maybe I'm being too optimistic here that big balance sheets can help the next time some piece of financial-engineering wizardry unravels. And this, of course, just takes us right back to NGDP targeting, with a fiscal backstop for bonus points.


    7. Well, you certainly shouldn't take me as the last word on it, but I do think, in recent history at least, reserves have not really been a constraint in US banking. And, yes, balance-sheet constraints would basically describe the same thing as capital constraints.

      It is different than in nonfinancial corporations, where you would expect cash invested in long term assets to be illiquid, and so firms can proceed with negative liquidation value as long as they can keep making payments, so liquidity is everything.

  5. This is the best discussion on IOR that I've found to date. Ken Duda's point is new to me and makes sense. Kevin E's point is a good one too (and matches my thinking until now). To the extent that IOR allows a larger base, maybe that's OK to a modest extent. But the current total reliance on IOR to maintain a much large balance sheet than would otherwise be seems to carry a large potential for mistakes. Right now I'd rather see the Fed let that balance sheet shrink without any more increases in the rate of IOR (including the last one).