Bernanke, breakevens, and volatility

I was surprised when I saw Tim Duy’s post, via Mark Thoma (Tim Duy is a blogger and economics professor at the University of Oregon) that took Bernanke to task for his poor management of market-based inflation expectations since the 2008 crisis.

This is what he says:

“Bernanke doesn’t appear to see that the inability to hold market-based inflation expectations at a consistent level as a problem:


What’s wrong with this picture?  Notice the volatility of expectations after the recession (Ryan Avent has made this point as well).  The Fed claims to have some mythical “credibility,” but it certainly isn’t evident in this graph.  If anything, it is clear that the Fed has failed miserably in establishing credible expectations for either 2 percent or stable inflation.  Instead, what they have created is very unstable expectations because of start-stop policy.  It is almost ludicrous to place so much blame on Congress for the unstable fiscal picture when they themselves are creating an unstable financial and economic environment.”

Psychological testing has shown than in building a story what matters most is not the completeness or quality of the information set, but rather its coherence relative to one’s priors. In other words, simple wins. And I’ll take this a step further: often the less we know, the easier it is for us to build a coherent story. (This is yet another well-kept secret of the money management business). The bottom line is that if a story is simple and intuitively compelling, it is fiendishly hard to disabuse people of it, no matter how false it might be. (Why do you think it was so hard for NBA talent scouts to give Jeremy Lin a shot?)

One would have thought that academia would be one of the spheres in economics best insulated from this confirmation bias. Evidently not.

I have three problems with his assertion.

The first is the premise.  The volatility patterns exhibited in his graph of 5-year and 10-year breakevens (the Fed’s preferred market-based inflation expectation input is the 5-year breakeven, 5-years forward, commonly denoted as 5y5y breakeven) is no different from the volatility pattern is ALL markets since the crisis. A good proxy for this is the VIX.


You will note that the pre-crisis average level of the VIX was 14.3 percent (green line), while after the crisis (using the same date used in the first chart) jumped to almost 23 percent. Moreover, as one might expect after a crisis, the volatility of the volatility also increased significantly.

In fact, as a market participant, I recall thinking in early 2009 how remarkably stable market-based inflation expectations were, given the prevalence of the fear of hyperinflation and overall levels of market volatility. You can get a sense of just how prevalent the talk of hyperinflation was if you look at the timeline of Google search for that word:


Second, at these levels of inflation, it is pretty clear that any volatility in market-based inflation expectations is the tail and not the dog. And the dog is doing all the wagging. What I mean is that variations in inflation at low levels have little negative effect on economic growth. At these levels the causation, empirically, runs from growth to inflation expectations, not the other way around.

We frequently hear in market commentary that every tick up in inflation is a tax on the consumer and every tick down is a windfall. But that’s not so much the case here at these levels of inflation. In reality, the level matters. A lot.

From low levels, upticks in inflation levels tend to indicate an increase, or an expected increase, in economic activity. And downticks, especially when growth is punk, are bad. The structural break in this relationship comes, again empirically, when inflation gets north of, say 5-8 percent. At this point the function that relates inflation to economic growth goes non-linear and the effects of inflation get very bad very quickly. So even if inflation expectations were more volatile, as long as they are well south of 5 percent it is not likely that they do much damage.

The third issue is conceptual. I could imagine someone looking at the chart of the VIX and the chart of breakevens and saying, okay, there was volatility in all markets, but that too was all Bernanke induced. And there may be some truth to that. But after a shock or crisis, normalization in markets typically follows the pattern of what is called disaster myopia: the acute memory of a disaster leads us to overstate the probability of its recurrence. In markets, that means volatility stays in the system. Volatility of the volatility stays too. Then, over time, as the memory fades, we progressively ratchet down our fears. (And, if a long enough time has passed, normalization eventually crosses over into complacency, FWIW.)

This is the pattern that you would expect to see across markets after a financial crisis like the one we had in 2008, and indeed, that is what is happening. Fear of the next Lehman, or the ‘next shoe to drop’ are waning, spasmodically, but waning. Hopes of a V-shaped recovery have exhibiting a progressively less amplitude as well.

Though there is much debate over whether the Fed has done too much or too little, it is hard to imagine that a radically different policy approach from the Fed would have led to a markedly different behavioral pattern, or to a rapid return to pre-crisis levels of volatility.

Leave a Reply

Your email address will not be published. Required fields are marked *