Many people think that the Fed should try to target inflation in “headline CPI”, the full consumer price index. Maybe they’re right (though I have some qualms). But maintaining this target over the course of several years is very different from hitting it precisely every time that data is collected. In fact, it’s entirely possible—and indeed, responsible—to target headline CPI in the long term while following a narrower price index in the short term.
Let’s consider an analogy. Suppose you find it prudent to target your annual spending at around $50,000 a year—your income, minus taxes and savings for retirement. In ordinary times, this works well, and it even works when it’s chopped up into shorter time horizons: say, a $12,500 target for every quarter.
You suddenly get appendicitis and go to the ER. A few weeks later, you find a $10,000 bill in the mail.
What do you do? You could say “well, I have this $12,500 quarterly target I need to hit, so I guess I’ll need to live on only $2500 this quarter. I’ll move back in with my parents and eat a lot of ramen.” But I don’t think many people would view this as a reasonable implementation of the consumption target. More likely, you’d say “my $12,500 quarterly target is still roughly accurate, but I’ll nudge it down to $11,500 to fill the gap in savings left by my recent episode, and account for the possibility of future medical expenses”.
It’s the same with inflation targeting. You’re trying to maintain the long-term rate of inflation at roughly 2% a year, or 0.5% a quarter. Then an oil shock comes and causes inflation of 2% in a single quarter. What should you do? You could say “I’m going to stick to my 0.5% quarterly inflation target, so I’ll tighten the money supply so that the other prices in the economy decrease by 1.5% this quarter to maintain the target”. But of course, most prices in the economy are sticky and won’t change in such a short time horizon—you’d need an extraordinarily tight monetary policy to get that kind of rapid deflation. And the macroeconomic consequences of that policy would put the financial crisis to shame.
It would be much more reasonable to say “Look, in the short term I’m going to exclude oil prices from my inflation measure, because they’re far too volatile to be part of a sane targeting regime. I recognize that if the price of oil relative to other goods increases consistently over time, this will induce some bias in my target relative to the full CPI*, and to correct for that bias I will slightly lower my medium-term target rate. This way I can anchor inflation expectations and maintain a target that works in the medium term, while sparing the economy from the wild monetary policy that would necessary to hit the target perfectly every quarter.”
Of course, there are other considerations as well—maybe oil prices carry some useful information about where the economy is headed, information that monetary policymakers should consider when making their forecasts. But that information is mixed in with information about the Chinese economy, the Japanese economy, the Indian economy, and whether the president of OPEC is going to wake up on the right side of bed in a few weeks. It needs to be interpreted in a very subtle way, not automatically forced to play an outsize role in monetary policy. And that’s what inflation hawks today don’t understand.
*Note: it’s not even clear why the long-term target should be the CPI, as opposed to (say) the GDP deflator, or the PPI, or some mix of all three. But that’s a topic for another post…