Inflation Expectations

Wondering what the masters of the AF universe think is the best market-based gauge of inflation expectations. go.

I don’t know if it’s the “best”, but there are actually inflation swaps that you can trade and that are actively quoted. Even if they are biased, they should be useful in measuring relative inflation expectations.

By ‘you’ - do you mean retail investors? If so, how and where?

Ok, I suppose retail investors would have a hard time trading these swaps. If you are just interested in quotes though, and have access to Bloomberg or a similar data service, there are live quotes compiled from various market maker contributors. There is a term structure of inflation swaps and everything.

I typically just look at the spread between 5 year Treasuries and 5 year TIPs.

If you want to look at things on a 5yr5yr forward basis, Fred has a series.

Cleveland Fed has a longer history of inflation expectations at a variety of horizons, but I think it’s only monthly.

^Breakeven rates are probably the most closely followed, but using it to actually predict where inflation is going is like trying to gauge a stock’s movement based on its PE.

I think inflation is harder to predict than interest rates, so many variables. But if I had to choose one thing to look at it would be the price of oil.

^ +1.

^and food and whoever does the calculations.

@StL If you’re looking out one month or three months, then you can actually be pretty accurate. It’s the longer-term inflation forecasts that aren’t that useful. Most people break down inflation into a few main components (core, food (as Neryblop mentioned), and energy (as you mention)) and then forecast each independently and build back to the headline forecast. Food and energy are typically strongly correlated with underlying commodity prices, but core is largely driven by macroeconomic factors. So the New Keynesian Philips curve approach will typically look at core inflation as a function of an output/unemployment gap plus some measures of expectations plus maybe some wage/unit labor cost info plus some lags. The importance of this is that any forecast of inflation in the future depends critically on the outlook for the economy. If you can’t forecast the economy, then you won’t forecast inflation in this framework.

Some people (market monetarists) focus on nominal GDP (NGDP). Monetary policy has control of NGDP (i.e. it can shift aggregate demand) through the growth of money, but the split between RGDP and PGDP is driven by the slope of aggregate supply. But then if you read them more, they’ll start questioning if PGDP is even a thing. It’s just some calculation we make to try to correct for stuff. So instead of looking at inflation, they’ll look at things like nominal wages that don’t rely on a bunch of assumptions.

Perhaps, but the only kind of inflation that matters is unexpected inflation. If economists could nail inflation three months out - or even one month really - on a consistent basis we wouldn’t get caught with our pants down. Instead history has repeatedly shown that we caught off guard by rising inflation.

^Maybe you could claim that only unexpected inflation matters for asset returns, but even then I could disagree (interest rates depend on expected inflation more than unexpected inflation, depending on how you measure it). What the Fed does is driven by inflation expectations, with all the consequences that has on the economy. Alternately, imagine an elderly person plans for retirement assuming 2% inflation, but then inflation expectations (not unexpected inflation, let’s say the Fed changes how they will do monetary policy) rises to 10%. Rates rise wiping out their fixed income holdings and then they have to deal with higher living costs as well. Expected inflation doesn’t matter then?

It seems like quibbling. I imagine you could rephrase everything I just wrote so that it is in terms of unexpected inflation. For instance, you could re-categorize the change in inflation expectations as unexpected inflation. But then we sort of divorce ourselves from what matters. Most everyone can agree what inflation expectations mean. They can’t for unexpected inflation.

Yeah, that’s pretty much what I was going to do. Unexpected inflation, either in real time or forecasted, is what causes shocks to the market. If inflation can easily be forecasted the Fed could accurately telegraph their interest rate moves. Instead we get unexpected inflation (or an unexpected rise in inflation expectations) causing the Fed to raise rates faster and higher than what had been priced in (among other consequences).

Maybe he means only unexpected inflation is relevant to changes in asset prices. Otherwise, it’s pretty far fetched to say that the expected level of inflation does not matter to anything.

historically, fed is typically a lagging indicator. she’s been yellen about it.

Right, expected inflation absolutely matters to the current price of an asset. Just like every other known (or best-guessed) variable, it’s already priced in. It’s unexpected inflation that causes major price swings.


I think we’re really just disagreeing over terms. I was operating under the definition of unexpected inflation as being the residuals from an MA(1) model on the monthly % changes in CPI. I think this is a common definition among economists, though not the only one. It is useful for thinking about what happens if prices change today, but not so much about prices in the future. I imagine to get unexpected changes in inflation expectations, you’d have to do the MA(1) for inflation expectations instead.