The Price of Credibility

A: So, diving into this paper, "Monetary Policy without an Anchor" by Bocola, Dovis, Jørgensen, and Kirpalani, it really crystallizes why central banks obsess over something called "de-anchoring" long-run inflation expectations.

B: It's the core fear, isn't it? If the public stops believing the central bank is committed to price stability, those long-run expectations float away, making future inflation harder to control. It becomes a self-fulfilling prophecy.

A: And this leads to a central trade-off. They highlight that central banks often accept pretty significant short-run output costs—like slowing the economy—precisely to secure that long-run credibility and keep expectations stable.

B: It's a signaling game. The key dynamic is the public constantly learning about the central bank's 'type'—whether it's a 'Hawk,' tough on inflation, or a 'Dove,' more tolerant of it—from every policy action. Each rate hike or cut is a data point.

A: We saw this play out in a very tangible way. Brazil's central bank, for instance, started raising rates in March 2021, explicitly citing these de-anchoring risks as a primary reason. It wasn't just about current inflation, but the future perception. So, how do you actually put a number on that credibility?

B: That's an interesting link. A sensitivity metric. But how does a surprise move in interest rates tell you anything about long-run expectations beyond its immediate effect? Isn't that just a direct policy impulse?

A: It's more indirect. A rate hike, for example, acts as a signal. It reveals the central bank's 'type' to an imperfectly informed public—are they a 'Hawk' who prioritizes inflation or a 'Dove'? That re-calibrates future inflation expectations, and *that's* what then helps moderate current inflation through the Phillips curve.

B: I see. So, the market isn't just reacting to the hike, but learning about the bank's true commitment. How do they disentangle that signal from all the other noise around a policy announcement? That sounds empirically tricky.

A: That's precisely where the empirical strategy comes in. They use high-frequency identification around policy meetings—looking at tiny windows of time to isolate the monetary policy surprises. And this signaling power, this reputational motive, it's strongest when the public is genuinely *uncertain* about the central bank's type, when its 'reputation' or ρ, is intermediate. If everyone knows you're a hawk, there's less to learn. So, what did the empirical data actually show for Brazil?

B: This is where it gets really interesting. They found a large negative elasticity: a one percentage point surprise in the Selic rate, for example, lowered 5-year, 5-year inflation compensation by almost half a percentage point, specifically 0.48 percentage points. For survey expectations, it was a 0.18 percentage point drop. That's a huge reaction!

A: Compared to other major economies, how does that stack up? Are we seeing similar sensitivities elsewhere?

B: Not at all. When they looked at the US and the Euro Area, expectations were much more firmly anchored. They reacted significantly less to those same interest rate surprises, which highlights Brazil's unique situation. This high elasticity then became a key target for calibrating the model to Brazil's economy.

A: So, the model, once calibrated to that high sensitivity, can then quantify the 'reputation dividend' you mentioned earlier. What does that mean in practical terms?

B: Precisely. It quantifies the benefit of having that strong credibility. Essentially, a central bank with high credibility can respond less aggressively to shocks because agents' expectations are already anchored. The big takeaway is that to replicate Brazil's policy trade-off without this reputational channel, the Phillips curve slope would need to be four to nine times steeper.