What are inflation break-evens signalling?
“It ain’t what you don’t know that gets you into trouble. It’s what you know for sure that just ain’t so. “ – Mark Twain
A closely held belief among central bankers and many market participants is that inflation expectations have a causal relationship with inflation and are a good barometer of it. It seems intuitive enough: players shift their behavior in anticipation of higher (or lower) prices and market participants have an incentive to create as accurate a forecast as possible.
With that in mind, it should be highly notable that market-implied inflation breakeven rates have declined and the discounted rate of inflation falls toward the 2% target over time.
Source: Stephen G. Cecchetti, Michael E. Feroli, Peter Hooper, Anil Kashyap and Kermit L. Schoenholtz. www. fivethirtyeight.com.
Except it’s not. An inconvenient reality is that there is not much theoretical soundness or empirical proof that inflation operates that way. In fact, changes to inflation expectations tend to be negatively correlated with changes in inflation (see: chart on right). The result is the same regardless of whether we use surveys of professional forecasters or the market-implied rate of inflation, and actually the latter might be slightly worse.
How wrong have they been? In the US, TIPS have only existed during a time of ‘well-anchored’ inflation but they have a tendency to under-estimate inflation (around two-thirds of the time) by around 70-80bps even over a short time horizon (6-12 months). The UK has a longer history with inflation-linked debt — since the 1980s. And while market rates outperformed survey-based measures in predicting inflation, the error was often on the order of 200-300bps over a 2 year horizon. Australia also has a more established time-series (since the early 1990s) and there too, the standard error was about 1.35% over a one-year horizon. In short — not very accurate.
Source: Survey of Professional Forecasters, Philadelphia Fed.
Both professional forecasters and the bond markets significantly under-estimated the rise in prices in the early 1970s and also significantly under-estimated the fall in inflation rates during the 1980s. Could the same be true today with a well-behaved return to 2%?
As previously noted on this site, some of best predictors of inflation is its recent trend (it has a high degree of serial correlation). Cecchetti, et al note that beyond this the main variables that have only minor predictive power are (1) money supply growth rates (2) credit growth and (3) changes in the value of the dollar. The first two variables are observable and operate with a lag, so I will dedicate a future blog post to analyzing the drivers of the dollar.
Post-script
This is not to say that inflation levels will not recede from 7% by the end of 2022. It is merely to highlight that we should understand that markets are (1) pricing in low levels of inflation (2) not very accurate at predicting it and this (3) has ramifications for other assets classes. There are a host of reasons for inaccuracy including liquidity premia, inflation risk premia (i.e. uncertainty), as well as hedging practices related to oil.