As an editorial disclaimer, this article is largely taking a “conventional” view on interest rate policy, as I am discussing a controversy that is happening between conventional economists (although my scepticism at various points should be obvious). Explaining how everything is incorrect from a post-Keynesian/MMT standpoint would require a longer text.
I have not been deeply concerned about the views of Kevin Warsh at the Fed, but he has caught attention as a result of returning the Fed to limited text on their policy statements. This is returning us a bit closer to situation in 1994, when the recently-deceased Alan Greenspan was Chair.
 |
| FOMC Statement, 1994. |
The above text still has more argle-bargle than necessary since the FOMC did not announce a target rate at that time. Market participants had to infer what the target rate was based on Kremlinology-style dissection of Fed statements, released data (such as monetary base statistics back in the Monetarist days), as well as what the open market desk was doing. This silliness was the result of the operational situation: excess reserves paid a 0% interest rate and the only way to directly borrow from the Fed had a stigma (the discount window). So the open market desk had to provide what it saw as the precise amount of excess reserves (which was tiny) to keep the Fed Funds market “in equilibrium” with other short-term interest rates (repo, Treasury bills). (This fit in with the “supply and demand determines interest rates” that neoclassical economists refuse to step away from, even when it has negligible importance within their own mathematical models.) There was uncertainty whether the Fed could keep the overnight rate near an announced target. Eventually, operating procedures got revamped and it was clear that the central bank could hit an overnight rate target (range), and so they just announce what the target range is. (Other central banks had borrowing/lending operations that were continuously used, so there was no worry about hitting target, as market rates were pinned by the central bank operational rate band.)
Forward Guidance
“Forward Guidance” is a bit of academic neoclassical nonsense that neoclassicals staffing central banks took seriously, and came to the fore during the Zero Interest Rate Policy (ZIRP) era.
Up until the late 1990s, people just thought about the level of overnight rates when thinking about central bank policy. At the extreme, it was possible to find people who felt that 10-year bond yields floated in space and had no relation to the overnight rate. However, during the “Bond Market Conundrum” era (Greenspan referred to the conundrum in a 2005 speech, and the topic stuck around for some time), people latched on to the fact that the rest of yield curve mattered. In particular, American consumers are highly sensitive to the 30-year conventional bond yield — not only do low rates help fuel house price bubbles, they get to refinance their mortgages at lower rates (which is largely unique to the American mortgage market). In the mid-2000s, bond yields were at the “wrong” level, and cancelling out the stimulus that the low policy rate was supposed to be providing.
Dropping the policy rate to 0% post-Great Financial Crisis (2008) was worse for conventional thinking, since that was allegedly the lower bound rate for interest rates (although we later had experiments in negative rates, but there was a limit how negative they could get without stresses hitting the financial system). Once you hit the “effective lower bound” for interest rates, how could the central bank stimulate the economy?
The economic fraternity then caught up to the fixed income mathematics that are embedded in standard financial models — as well as their own neoclassical models. What the overnight rate was doing was not the only thing that mattered — the entire risk-free yield curve expressed the reaction function of the central bank, and that is what mattered for economic expectations (which allegedly drive the economy). If they wanted to stimulate the economy, they needed bond yields to be lower, as bond yields reflect the expected path of the overnight rate. (They also were mechanically related to the interest rates that entities borrow at, such a 30-year conventional mortgage rates. This means that the entire curve matters, even if you think “reaction functions” are a load of hooey.) Although they did launch “Quantitative Easing” (QE) to attempt lowering yields (and it didn’t really work, as even a glance at standard financial/economic models would tell you), they needed to “jawbone bond yields lower.” That is, they needed to convince bond market investors that they would not hike as aggressively as was priced into the markets. Since it costs Fed officials nothing to be wrong about future rate hikes while it is extremely painful for wrong-footed bond investors, just saying “bond yields appear to be too low” was not enough. Instead, they made pinkie promises that they would not hike over a certain time horizon. Although this in theory could work, it also costs nothing for the central bank to break said “promise,” which puts the market value of said promise as $0. (See the Appendix on how central banks can influence bond yields.)
Warsh’s Reversion to Past Form
Reducing the text bloat of Fed statements is not going to change much without a greater overhaul of communications strategies. The micromanaging of the text within statements (which now receives the Kremlinology treatment, since central bank watchers need to look busy) is just one method of signalling to markets whether Fed officials think market pricing is out to lunch or not. They would also need to clamp down on other signalling methods to truly get away from the “forward guidance” mindset.
Although this might be attempted, it is only going to work for as long as Fed officials are not panicking about the economic outlook, one way or another. (It would also require Fed Governors to not leak discussions to selected “friends,” which is somewhat of a sore point at the Fed historically.) Based on the post-1990 dataset, they could go for up to a decade without hitting a panic mode, but a panic will happen, and a shift away from “open mouth operations” would almost certainly be reversed.
Why Not Just Follow the Markets?
Warsh might have bought into some free marketeer nonsense about letting markets determine rates. Free marketeers have a hard time grasping that the overnight policy rate in a fiat currency is an administered monopoly. Unfortunately, the people pushing theories about “market-determined interest rates” the most are parochial Americans who attempt to generalise the pre-Fed American experience without realising that America was an “Emerging Market” tied via a currency pegs to senior currencies. American rates were market-determined spreads over the policy rate in the senior currency (mainly the Bank of England’s Bank Rate). Spreads are market-determined, as is the case now in emerging markets with currency pegs or spread markets within developed markets. Once you have broken any pegs for your currency, your central bank has no hard reference point for its policy rate. (As a counter-example, the Bank of Canada periodically suggests that it shadows the Fed, but that just reflects the ideological biases of the staff, and the policy rates can diverge in practice. However, the large economic linkages from the pre-Trump Free Trade era ensured that economic cycles were largely synchronised.)
Once we accept that the Fed is a monopolist setting the overnight rate, it is clear that it is silly to think it can follow “market rates.” Market participants are guessing how the central bank will set the rate in the future. Blindly following what is priced into the market is just going to degenerate into a free-for-all where some participants attempt to manipulate short maturity rates. It also is a statement that the entire analytical apparatus at the Federal Reserve is worthless — it suggests that the battalion of Ph.D.’s employed by the Fed have no idea how to model the economy. Although that might very well be true, it is a view that will not sell well in the Jackson Hole confab (which means that it would not survive institutional resistance).
Inflation-Targeting Central Banks More Sensible
One of the institutional advantages of having a formal inflation target is that it puts central bankers on the spot — they have a clearly defined objective, so they better come up with a good reason why any misses happen. The institutional response is regularly publishing inflation/economics outlooks. The forecasts can be wrong, but they are forced to explain what the uncertainties are within the report (and associated press conference).
The Fed’s communications strategy was converging towards this (dot plots, etc.), but the issue is that the Fed takes advantage of having a squishy mandate and top officials decide how to hold themselves accountable. Announcing “I will set the policy rate were market signals indicate” will make market Austrians swoon, but one would get obliterated by the legislative body that set an inflation mandate when they ask “And will that guarantee that the inflation target would be hit?”
Admittedly, this could be cover for Warsh just setting interest rates where President Trump wants them (as he reiterated on Wednesday, the answer is “lower”). However, given that President Trump is currently losing control of the nation’s capital to algae, it is unclear that he has the political capital to take on the Fed.
Summary
Conventional central banks set the overnight rate based on their outlook for the economy (particularly inflation), regardless of the existence of a formal inflation target. Financial conditions matter, but financial crises are not that frequent in the modern world (roughly once per decade).
Central banks cannot ignore the rest of the yield, particularly the Fed due to the preponderance of fixed mortgages. Sooner or later, bond yields are going to be at what policymakers think is the “wrong” level, and they are going to prefer that bond yields move rather than disproportionately move the overnight rate to compensate.
Unless the central bank rate setting committee cuts off all communications with the outside world, they are going to communicate their unhappiness with the level of bond yields somehow when the time comes.
None of this really matters one way or another for financial asset markets (other than money market instruments). Although people love pinning explanations for market movements on the Fed, they only really matter in a crisis.
Appendix: How Can a Central Bank Direct Bond Yields?
Neoclassical economists are fixated on two wrong ways the central bank can “set” bond yields.
Supply and Demand: by buying/selling a quantity of bonds, the magic of supply and demand curves will result in measurable bond yield changes. This is arguably true for the extremely thin ultra-long (20+-year maturities), as there is negligible private sector supply of similar duration instruments. (Conventional 30-year mortgages are amortising and have pre-pay options, so the effective duration is often closer to that of a 7- to 10-year Treasury.) Once we are in the belly of the curve, there is so much duration supply that the Fed cannot hope to measurably influence it. At best, they can affect term premia, which are not that large when compared to potential changes in the expected rates (unless the term premium model has gone off the rails).
Some of the more extreme academic economists take DSGE model behaviour literally (Market Monetarists being a key example), and they think the central bank can achieve practically anything by announcing a “credible” policy. This thinking showed up in the promises about not raising rates for N months. Unfortunately, neoclassical economists really did not take the “time inconsistency” problem for monetary policy (it apparently only matters for fiscal policy). It costs central bankers nothing to break a previous commitment, so that is exactly what they will do if they think it is necessary.
The answer is to drop “supply and demand curves” and “reaction functions,” and do the obvious: price signals. Set a target for bond yields, and intervene if they get out of line. There is historical precedent for this (operational framework that developed during World War II), and was discussed by Keynes in his General Theory. To the extent that the central bank is credible, just the target might be enough to keep bond yields within a target band (much like currency pegs). They would only need to step in if markets are convinced they are too far out of line.
No comments:
Post a Comment
Note: Posts are manually moderated, with a varying delay. Some disappear.
The comment section here is largely dead. My Substack or Twitter are better places to have a conversation.
Given that this is largely a backup way to reach me, I am going to reject posts that annoy me. Please post lengthy essays elsewhere.