David Beckworth’s excellent Mercatus podcast series has come out with an interview of Neel Kashkari, who is President of the Minneapolis Fed. (The same interview includes an extensive discussion with vice president Ron Feldman as well, something I’ll cover in a later MoneyIllusion post.)

Kashkari is one of my favorite people at the Fed, and I agree with much of what he has to say. As usual, however, I’ll focus on the few points of disagreement. Here’s Kashkari:

When I travel around my district and I talk about the Fed’s current framework, I get a lot of pushback on our 2 percent target. I can only imagine the outcry if we were to try to raise it to 3 percent or 4 percent, number one.

Number two, so far, we can’t even hit our 2 percent target. If we announced a higher target, 3 percent or 4 percent, or a base-level target, it isn’t obvious to me why anybody would believe us or should believe us.

I don’t think this is right. Let’s start with the phrase “we can’t even hit our 2 percent inflation target”. What does that actually mean? It might mean:

1. We are incapable of raising inflation to 2%, because monetary policy is ineffective.
2. We can control inflation, but our current procedures lead to a consistent bias towards slightly under 2% inflation.

I strongly believe that the second interpretation is correct; indeed if the Fed actually were incapable of pushing inflation up to 2%, then we’d probably be currently experiencing deflation. Instead inflation is quite close to 2%, and the Fed is steadily raising interest rates. They are obviously not “out of ammunition”.

Kashkari’s a smart guy, and certainly knows this as well. So I think he and I probably share the view that it’s the second interpretation that is correct. But what does that imply about a higher inflation target?

Suppose Fed procedures lead to inflation persistently running about 0.3% below the target (perhaps because they treat it as a ceiling, not a symmetrical target). Then an increase in the inflation target from 2% to 3% or 4% would raise the actual inflation rate from 1.7% to 2.7% or 3.7% (assuming the bias stayed the same). That’s not my preferred policy, but it’s certainly something the Fed could do. There’s no reason to suppose a higher inflation target would not be credible in a general sense; indeed why would the Fed announce such a target if they didn’t intend to raise the actual rate of inflation up closer to the new target? Yes, 3.7% inflation is not 4%, but it’s much closer to 4% than is 1.7% inflation.

I happen to oppose raising the inflation target, but not because I don’t think the Fed could hit it, rather I don’t think they should be targeting inflation at all. And if they insist on doing so, I’d prefer something closer to Bernanke’s recent price level targeting proposal.

I think we haven’t decided yet how small a balance sheet we want to return to. I think there are advantages to both a corridor system and a floor system. Obviously, the Fed had a corridor system for most of its history, up until recently, when we adopted the floor system around QE. We think, on the margin, a floor system is somewhat easier in terms of operational complexity, a little lower complexity.

I also see an advantage that, if we are in a low r-star environment, where we could be hitting the zero lower bound more frequently, and we might have to turn to QE in future downturns, then if we’re already in the floor system, admittedly with a smaller balance sheet, maybe it’s somewhat easier to then ramp up QE, as opposed to having to shift from a corridor system to a floor system in the future.

So on the margin, I think there are some benefits to a floor system, but I see advantages both ways.

Here Kashkari is referring to the fact that a floor system involves interest on bank reserves (IOR). This was adopted in 2008, as a way of allowing the Fed to do massive QE without reducing interest rates to zero. Without a floor system (i.e. without interest on reserves) a large QE could drive interest rates all the way to zero. (Of course it could also create hyperinflation, it depends on where the economy is when the money is injected.)

I’m not convinced by the argument. If QE is likely to lead to hyperinflation, then don’t do it. If QE is likely to push interest rates to zero, then you probably want interest rates to fall to zero.

Consider the fall of 2008. If the Fed had done QE without a policy of IOR, then interest rates would have quickly fallen to zero, many months before they actually fell to zero (actually 0.25%). But in retrospect this would have been good. Indeed Bernanke noted in his memoir that monetary policy was too tight after Lehman failed in September 2008, and although he did not say this, it was IOR that allowed policy to be tighter than desired, given that the Fed was doing QE to prop up a shaky banking system.

There is an argument that IOR gives the Fed one more degree of freedom, one more tool to use. But I fear it leads to the Fed trying to do too much—aiming policy at both the banking system and the broader economy. I’d like them to focus like a laser on aggregate demand (i.e. NGDP growth) and not having IOR as a tool makes it more likely that they will do so.

In plain English, if you are doing massive QE to rescue the banking system, it’s a good bet that monetary policy is too tight. The last thing you want to do in that case is make it even tighter by paying banks an above market interest rate to sit on their bank reserves.

And by the way, unless I’m mistaken, the Fed’s policy of paying banks an above market interest rate is also illegal. The Congressional authorization of IOR mandated that the rate not be higher than market rates. But it is higher. (Please correct me if I’m wrong on this legal point.)

PS. There’s much more that could be said about IOR, and I refer readers to the excellent work of George Selgin, David Beckworth, and others.

PPS. The following two graphs show the system as it was before IOR was instituted in 2008, and as it looks when IOR sets a floor under interest rates.

Screen Shot 2018-02-06 at 2.06.21 PM.png
Screen Shot 2018-02-06 at 2.06.31 PM.png