Rates: Why the yield curve should do its own thing

Yield curve control: three simple words. And a simple concept, at least it should be. But it’s very nuanced.

The precise wording suggests that the entire curve is controlled. Control here implicitly means a cap on yields, and so is a means to preventing yields from rocketing higher. It is executed through the central bank standing ready to buy bonds should the market yield drift above the desired yield. Straightforward enough, but why do it? And on what tenors? And what are the potential unintended consequences?

A starting point is to note that yield curve control is all about the price.

The supply versus demand for bonds typically determines their price. Should a central bank wish to control that price (and by implication it’s yield), it will stand ready to buy those bonds. So in the case of yield curve control, we know the price, but not the quantity. Contrast that with quantitative easing where the central bank knows the quantity it will spend, but has no target price for the bonds it buys. The object here is to add reserves to the system, that can be deployed in the wider economy.

In the case of YCC, we know the price, but not the quantity. Contrast that with QE, where the central bank knows the quantity it will spend, but has no target price for the bonds it buys

Yield curve control feels similar to quantitative ease, as it too adds reserves. But the object of the exercise is quite different. It is far more about containing and controlling market rates. The Federal Reserve deployed this policy during the post-war years. The Reserve Bank of Australia currently employs a policy with a concentration on the 3-year tenor, and the Bank of Japan has had a policy concentration on the 10-year since 2016.

In all cases, one key objective is to provide some certainty for funding costs for the government and the wider economy. The Fed is now considering something similar.

The dominant view centres on controlling the front end, starting with the 2/3yr area. These are both auction maturities for the US Treasury, mapping out the first couple of points on the curve that extends to the 5yr, 7yr, 10yr, 20yr and 30yr benchmark maturities. The advantage of a front end focus is that it is more controllable.

The Fed sets the funds rate with certainty, and in that sense has far more control of the 2yr rate than the 30yr rate. The latter has a much longer nose into the future and is thus more heavily influenced by longer-term interest rate and inflation expectations.

Therein lies a risk for a pure front end focused policy - the risk that the curve steepens from the back end; there is nothing to stop 10yr to 30yr yields from shooting higher. This may not be a bad thing as it reflects a solid reflection of optimism on the future for the economy. But it could be damaging if the expectation was misplaced, as it means higher than ideal funding costs for government, corporates and the personal sector. Another driver could be supply, as the US Treasury has had to increase issuance for Covid-19 impacted financing requirements. Countering that impact has benefits.

Therein lies a risk for a pure front end focused policy - the risk that the curve steepens from the back end; there's nothing to stop 10-yr yields from shooting higher. This may not be a bad thing. But it could be damaging if the expectation was mis-placed.

The argument against yield curve control for longer tenors are twofold.

First, it is deemed to be more difficult to control longer yields as they are more slavish to longer-term expectations. Second, artificially fixing such long tenor yields robs both the market place and the Fed of an important discounting function that the shape of the yield curve and the level of long rates provides. A counter-argument is only by controlling long tenor rates can economy-wide rates be truly contained, ranging from corporate refunding in the 5-10yr maturities and the likes of 30-year mortgage rates.

There is one important technical argument in favor of longer tenor yield control to do with the required quantities required to achieve such control. The thinking here is most of the market capitalization is in shorter tenors, so to control this, bigger volumes would need to be bought by the Federal Reserve. On top of that, there is bigger bang for the buck in longer tenors by virtue of the fact that these are longer duration product, so a small effect on price would have a bigger effect on yield. In other words, the Federal Reserve may be able to spend less to control yields in longer tenors.

If the Federal Reserve set 1% as a target for the 10yr, that is entirely defendable with the funds rate at zero, and a curve effectively mapped out as 100bp. Something similar could be said for a 2% target for the 30yr

But that in part depends on the level of yields chosen. If the Federal Reserve set 1% as a target for the 10yr, that is entirely defendable with the funds rate at zero, and a curve effectively mapped out as 100bp. Something similar could be said for a 2% target for the 30yr. For levels like these to be challenged there would need to be a material breakout in inflation expectations to the upside. The front end is more straightforward in the sense that the Federal Reserve has complete control over the funds rate, and in that sense could set 25bp as a viable target for the 2yr yield (versus the funds rate of zero-25bp).

There is another strategy that the Fed could pursue. It could announce a set target rate for the 2yr (or 3yr) and vow to cap that yield. And then for longer tenors it could assert that there will be non-specific yield curve control employed. In other words, the Fed could map out a level for yields for longer tenors that they would object to, setting implicit caps to yields right out the curve, but without announcing what these cap levels are.

This more fluid policy would be easier to deploy, as the targets can be moving ones. The downside is it leaves the market guessing as to when it hits those caps, and in that sense subject to excessive conjecture.

So what does this all tell us?

First, there is no certainty that the Fed will deploy yield curve control. Right now there is no specific need to. The 2yr is practically anchored near the ceiling of the zero-25bp funds rate range, and the 10yr is still below 1% (and the 30yr below 2%). Moreover, we observe that the 5yr is rich to the curve on the 2/5/10yr fly, which is a signal that this is not a bear market for bonds at this point.

There is another strategy that the Federal Reserve could pursue. It could announce a set target rate for the 2yr (or 3yr) and vow to cap that yield. And then for longer tenors it could assert that there will be non-specific yield curve control employed

Second, if the Fed does venture into yield curve control they would have to make the call that yields were threatening to obstruct funding circumstances, that could, in turn, threaten the recovery. It seems likely that they would start off by setting a cap on the 2yr. That would limit the money market curve. But steepening pressure could build from the back end in consequence. Forward guidance would be of minimal use here, as the Federal Reserve simply can't provide this with certainty for 10yrs, and investors know that.

Third, the only way to contain the curve would be to deploy a bond-buying out the curve with the objective to shepherd the curve along a tolerable range. The main trigger points here would be soft caps, where soft buying would happen as a tolerance range was entered, morphing to stronger buying as the ultimate cap rate was approached. The range and the cap rate would ideally be known by the market place but could remain unannounced.

That would give the Fed the flexibility to change both the range and cap, depending on wider circumstances.