Home > Bond Market > Imagining the Unimaginable

Imagining the Unimaginable


Last week I met and spoke with some bright minds at a big reinsurance company, who were sampling some views on inflation. Among the questions, however, were ones concerning my views on nominal interest rates, to which they are more directly exposed.

Since I was a rates strategist long before I was an inflation specialist, I do have some opinions on the matter.

Early this month, I wrote an article asking “which consensuses are worth fading?” in which I noted that of all of the “consensus” views, I am most sanguine about the view that interest rates will rise over the course of this year. Now, there are lots of ways that this view can be derailed. For example, there are a lot of concerns about a slowdown in China and what that might mean for global growth. There are other land mines, such as the risk of a default of Puerto Rico, which could send investors scurrying for at least a short time into nominal bonds. And, of course, we are not out of the woods ourselves, and a dovish Chairman Yellen (if in fact she turns out to be as dovish as we all expect) could easily stop or reverse the taper even though that does not seem to be the plan at the moment according to the Wall Street Journal’s Jon Hilsenrath.

But abstracting from the chance that a metaphorical meteor might strike the earth and ruin all of our plans, what are the chances of somewhat higher rates, or drastically higher rates? In my mind, the chances of these different outcomes derive from the types of causes that could provoke them.

10-year rates at 4% by year end – To get 10-year rates to approach 4% (a level last touched in 2010 and not seen on a closing basis since before the crisis in 2008) doesn’t require a miracle. The Fed is in taper mode, and there reportedly remains a pocket of ‘negative convexity’ that could turn a mild selloff into a major selloff if interest rates rise towards 3.25%. It was the ‘convexity trade’ that helped fuel the move in 10-year notes to 3% last year (see my comment about that here) from 1.65% in May, and another convexity-triggered selloff could easily cause rates to reach 4% at some point this year. Of course, that would still be an interesting technical development, since it would be the largest deviation above the log-channel lower in rates that has been in force for more than thirty years (see chart, source Bloomberg).

channel

10-year rates at 5% by year end – Five percent 10-year rates seems outlandishly far away, and we haven’t seen them since 2007, but we should recognize that this is roughly a neutral nominal rate. If real growth is expected to be 2.75% on average over time, and inflation is expected to be around 2.25%, then r + i = 5%. If the Fed is normalizing policy, is it really that much of a reach to get normal market rates? I don’t see 5% as being outrageous. However, realistically I would have to say that there will be a lot of friction between here and there, by which I mean that as interest rates rise, investors will find them increasingly attractive and will rotate from equities to bonds. That will make a 200bp selloff somewhat difficult, in my view. But against this, we need to keep in the backs of our minds the possibility that the Federal Reserve could choose to start selling securities from its portfolio at some point; while the Fed professes to be relying on its reverse repo facility to be able to drain liquidity as needed, that’s only plausible if they need to drain relatively small amounts of liquidity (tens or scores of billions). As rates approach 5%, losses in the Fed’s SOMA portfolio will be large enough that it will be technically impossible for it to fully drain all of the reserves they have added – and will be a political football, no matter how the Fed chooses to account for a mark-to-market loss (see my article from a year ago on this topic here and a follow-up article with additional issues here). I am not making any predictions about what the Fed will do or not do when rates start to rise past 4%; I only point out that there will be a lot of zeroes involved and that tends to affect decision-making. A move to 5% isn’t, in short, completely crazy although I don’t think we’ll get there.

10-year rates at 8% by year end – How can you get really ugly outcomes, like 8% nominal rates (which we haven’t seen since 1991)? This is outside the realm of forecasting. A 500bp move in a year is roughly a 4-5 standard deviation event. In the post-WWII period we have never had a 500bp move on a year-end to year-end basis. In fact, we have never had 10-year rates move more than 400bps in a 12-month period. So, this is really outside of the range of outcomes one could reasonably expect in a normal world.

This is, of course, not a normal world. But it is non-normal because weird departures from normality happen stochastically and, when they do, the distribution we draw market outcomes from is unknown. Put another way: for rates to rise to 8% in a year would take something really crazy. So, we can’t make predictions, but we can play with entertaining suppositions and I will do that in a moment. But before I do, I just want to make very clear that guessing how rates would come unglued and get to 8% in 12 months is, since it relies on a chaotic break, probably unknowable in advance. We can, though, test the limits of imagination to see if we can come up with a plausible scenario in which such an outcome would not be impossible.

And here is where inflation, and specifically inflation expectations, come in. The dynamics of nominal interest rates imply that at low levels of nominal rates, movements are caused mostly by changes in real rates (thus the high beta of TIPS at low rates) while at high levels of nominal rates, movements are caused mostly by changes in inflation expectations.

Suppose that inflation expectations can be characterized as “multi-equilibrium,” meaning that they are mean-reverting within some ranges but then can jump to a new equilibrium when expectations become “unanchored.” I’m not particularly enamored of that notion, because it has been used to conceal of a lot of bad econometrics, but let’s just suppose it’s possible that inflation expectations can both anchor, and become unanchored. We could hypothesize, for example, that consumers (and investors) don’t encode “1.987% inflation” or “3.5093% inflation” or “6.421% inflation,” but rather “low inflation,” which means anything where inflation doesn’t enter into daily consideration, or “medium inflation” (which is where inflation considerations cannot be overlooked), or “high inflation” (which is where inflation considerations are the prime concern).

If that describes the way that inflation expectations behave – and I think it is fair to say that, at least, it is the way that financial journalists behave – then it’s plausible to say that inflation expectations might move very rapidly from a distribution centered around, say, 2% to one centered around 5%. And that, in turn, could trigger a very sharp move in nominal rates. If that happened, it could plausibly be worse than historical precedents if only because the system is far, far more leveraged now than it was in the late 1970s, when we last saw a sharp ramp-up in expectations.

Again, none of the foregoing is a forecast per se, but a statement of possibilities. I expect nominal rates will at some point this year (probably in late Q3) approach 4%, and I think there’s a measurable chance that things could get ugly enough, in an environment where Wall Street dealers are discouraged from providing liquidity by leaning against the flow, to push rates towards 5%. I don’t really think that’s very likely, though. And I think it’s quite unlikely that rates could approach 8% this year, or even next year. But if you’re thinking about tail risks – and you should be – it’s less important that it may happen than that it can happen. The point is not to try and look for the signals that this particular scenario is unfolding; by the very nature of such a chaotic move, it is very unlikely that we’ll correctly guess in advance what will actually cause it. But, if we can imagine a not-wickedly-outlandish scenario in which this outcome can be achieved, then it means the unimaginable is no longer unimaginable. It is possible, and the next question is whether it is worth hedging against that tail risk.

Advertisements
  1. eric
    January 21, 2014 at 8:30 pm

    Very thought provoking post.

    I’m curious how you transpose these scenarios into the 10-year TIPS domain.

    I’m assuming you think the odds of 1.5% on TIPS are similar to your first projections, but that the tail risk of TIPS going up more than 200bps is virtually nill. But in your second projection, where do TIPS go? Closer to 1.5 or closer to 2.5?

    • January 21, 2014 at 9:22 pm

      I would say that if historical trends hold, 4% 10y notes would be about 1.4% on TIPS; 5% 10y notes would be about 2.2% on TIPS (that is, most of the move would be real rates).

      For the record, at 8% nominal rates, real yields would be about 4%. Obviously there are some big error bars. And on my model, 10% nominal would only give 4.3% real yields.

  2. Elliot Royce
    January 21, 2014 at 8:50 pm

    I like your multi-equilibrium construct — life does not move in a straight line. My own view (for what it’s worth) is that we are in a long-term secular trend that will place a premium on the guaranteed return on bonds (particularly TIPS). This trend is the result of the baby boomers retiring coupled with the projected low returns on equities. The lateral thinking here is that rather than going for the least dirty shirt (dividend paying equities at 2% yield) many investors will actually want to lock in a 1-2% real yield via TIPS. While I’m less certain about regular bonds, I suspect that retirees will be attracted to their yields given that inflation is likely to continue low for a long time until, poof, inflation is no longer low. As we’ve seen, inflation never proceeds as a trend: I lived through the late 1970s and early 1980s and no one saw that coming.

    • January 21, 2014 at 9:25 pm

      Exactly. I was talking to someone the other day and they remarked that Goldman’s strategist forecast was for a “slow rise in rates and inflation.” And it occurred to me that that’s what ALL strategists using models will typically say, because models aren’t good at capturing the multiple-equilibrium/messy break concept. And this is obviously also true at the Fed. There is no model at the Fed, I can almost guarantee, that even contemplates 4% inflation next year. Much less 8%. And yet…that’s not exactly unprecedented. They certainly won’t see it coming, and later they’ll go make up some reason that their model worked, but was just missing a little parameter.

  3. eric
    January 22, 2014 at 11:53 am

    I also think one shouldn’t rule out the possibility of returning to 1.5% nominal. Here’s a possible catalyst:

    http://blogs.cfr.org/geographics/2013/12/04/greeksurpluses/

    • January 22, 2014 at 12:13 pm

      there are certainly AMPLE numbers of potential calamities out there! Greece is probably on the small end by now. But what about Spain? I don’t disagree there is such a risk.

      ________________________________

  1. No trackbacks yet.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: