Front end trades are a bet on the calendar, not the economy

Short duration traded at 81.12 this morning, -0.01% on the session. I want to strip some mystique off the two-year, because it is a more mechanical instrument than the commentary implies.

It is mostly arithmetic

The two-year is close to the average expected policy rate over the next eight quarters, plus a small premium. That average is set by a finite, scheduled list of meetings, each with a small number of possible outcomes.

So a view on the front end is a view on that schedule. Not on growth in the abstract, not on the labour market as a theme, but on what happens at named meetings on known dates.

What I actually do

• Write out every meeting to the two-year horizon
• Assign a move and a probability to each, honestly, including the ones I have no view on
• Average it, add a premium, and compare to where the market is
• Only act if the gap is larger than my own error bars, which it usually is not

The discipline is in that last step. Most of the time the answer is that the market has it about right, and the honest response is to do nothing.

Why the front end punishes narrative

A recession call is not a front end trade until you can say which meetings it changes.

Plenty of people have been right about the economy and lost money in the front end, because the timing of the policy response did not match the timing of the view. The instrument does not pay out on being correct eventually.

Where I am

Flat, because my arithmetic and the market's land close enough together that the difference is inside my own uncertainty. That is not a failure to have a view. It is the view.