Measuring Term Premia (And Other Comments)

I have been slowly getting back into things, and been thinking about what to write next. Based on a conversation with Gabriel Mathy on Twitter, I just want to quickly discuss how term premia ought to be measured. I think I have mentioned this before, but I felt like running through the arguments briefly again. I then have a few other unrelated comments to fill out this article.

How to Measure Term Premia

In order to keep this brief, I will assume that the reader is familiar with the concept of the term premium. I have a pair of useful background pieces.

1. A primer that was one of my earliest articles on my blog.

2. A discussion of how the term premium should be approached. To recap, how you use the term premium partly determines how you should measure it. I think the points I raise in this article are not appreciated enough.

I am going to assume that the reader is approaching the term premium like an economist: how much of a bias is there in forward rates versus market participants’ unbiased expectations for the path of the policy rate? (A positive term premium raises forwards relative to those unbiased expectations.)

One popular way of estimating the term premium is to compare surveys of rate expectations of economists to market yields. (The methodology is quite complex in modern academic models, but the basic principle is there.) This has two problems.

1. Economists are not the ones taking interest rate risk. The economist profession in finance is a separate group that go by their own rules, and are subject to things like herding. Interest rate calls can be subordinated to what the equity team wants, because equities bring in more revenue.

2. Economists quite often work in a committee structure, and so there is a decision lag. From discussions I had with market participants, the suspicion was that economists’ interest rate calls lagged market moves.

3. The previous factor plus the reality that economists update views based on things like publication schedules means that we have a hard time aligning survey dates with the dates for bond yield observations.

The time alignment difficulty is a serious problem when we consider that bond yield trends are driven by largely instantaneous moves in response to news. A few days lag means that the market yield is non-comparable to a survey based on information before the arrival of the news.

My feeling is that academics are not bothered by this since they are applying a mental model they have taken from equities. The “fundamental” news for equities (published earnings, analyst surveys) are low frequency, so the movement in equities is explained as a result of high frequency changes in risk premia.

This makes no sense for the term premium (for reasons that are seen in my linked articles). Unless something radically changes, the term premium ought to be largely stable. As such, instead of using a survey to measure rate expectations, we need to do a survey of term premium estimates (which are a function of maturity).

One problem with such a survey is that asking economists to respond would be pointless, and they are the ones who normally respond to such surveys. Real money investors would probably not be happy to respond to nosy enquiries about valuation metrics. My guess is that the only way to have a continuous survey would be to have primary dealers respond. Since they need to have an ongoing working relationship with the central bank/Treasury, they could get arm-twisted to take a monthly survey at least slightly seriously.

The exact form of the question might be awkward. Is it what the respondent thinks what the term premium (at a given maturity) ought to be? Is it what the respondent thinks the market “consensus” is?

Inflation Book Woes

I had largely finished a first draft of my “What is Inflation?” manuscript. The idea behind the book is that it is a primer that is not in any sense an attempt to forecast inflation. (Since I want to sell my books on multi-year horizons, forecasting would have to be extremely long term.) However, it was extremely awkward having the book finish with all the inflation charts spiking at the very end. Since the recent experience is the first time inflation has been interesting in most of the developed world in decades, it would be strange to completely ignore the spike.

I was hoping that the spike would settle out to more sensible levels, and I could just discuss the spike and leave open the question whether it would head back down to 2% or 4% or whatever. The problem is that the darned spike is just not stopping.

My plan is to try to give an overview of the pandemic spike, make a few educated guess about the war spike, and leave it at that. I hope to put out the drafts in the coming weeks.

War Comments

People and markets are reacting to “peace deal” news, which I think is mistaken. Barring some unlikely scenarios, a peace deal is only going to happen after some form of resolution of the battles in the south and east of Ukraine.

The Russian advance towards Kyiv has been smashed, and so extravagant war aims by the Putin regime have to abandoned. What is going to be determined are the borders under Ukrainian control. Right now, the Russians are on a lot of territory that the Ukrainians are not going to give up without barring a total military defeat. The Ukrainians are attacking Russian rear areas, and so the Russians will have to give up ground. Until that process advances to a point that the Russians have a defensible front line, neither side will agree to a ceasefire. I would guess that it will be at least a month before the military balance is clearer, at which point the serious ceasefire negotiations start.

Email subscription: Go to https://bondeconomics.substack.com/ 

(c) Brian Romanchuk 2022