Lecture 1A: Using the Taylor rule in practice

Start here if this is new to you

In the previous chapter the rule looked exact. In practice three of its ingredients are choices, not facts: which inflation number you use (headline or core, this year's or a forecast), what the neutral real rate r* is (nobody observes it; it is estimated, and estimates disagree), and how you measure the output gap (potential output is also an estimate). Change one choice and the prescribed rate can move by a full percentage point or more. That is not a weakness of the exercise; it is the point of this course: you learn to make those choices explicit and defend them with data.

Exercise: where "should" the ECB's rate be in September 2026?

Data: HICP inflation 3.3%, core about 2.5%, actual DFR 2.50%; assume an output gap of 0.

Choice	π	r*	Prescription
Taylor (1993) as is	3.3	2.0	2 + 3.3 + 0.5(1.3) = 5.95%
Lower r*	3.3	0.5	0.5 + 3.3 + 0.65 = 4.45%
Core inflation, lower r*	2.5	0.5	0.5 + 2.5 + 0.25 = 3.25%

The answer moves by almost 3 percentage points just by changing assumptions. Which inflation measure, which r*, which gap? Those choices are the heart of the course. (These are illustrations, not official estimates.)

What you can change in the rule

Inflation measure: CPI, PCE, core, HICP, or forecasts of inflation.
Neutral rate r*: fixed at 2%, or estimated with models (Laubach-Williams, Holston-Laubach-Williams). Recent estimates are often much lower than 2%.
Activity measure: the GDP gap, or the unemployment gap (how far unemployment is from its normal level).
Weight on the gap: 0.5 or 1.0.
Interest rate smoothing: central banks move in small steps instead of jumping straight to the prescription: i_t = ρi_t−1 + (1 − ρ)i_t^Taylor, typically ρ ≈ 0.85. Example: last rate 2%, Taylor prescription 6%: i = 0.85 × 2 + 0.15 × 6 = 1.7 + 0.9 = 2.6%. The rate moves only 15% of the way each period.

Why central banks move in small steps

If the rule says 6% and the rate is 2%, why not jump to 6% at once? Because the inputs are uncertain and get revised, because a sudden jump can unsettle banks and markets, and because moving gradually lets the central bank watch how the economy reacts. Smoothing captures this: each period the rate moves only part of the way (with ρ = 0.85, fifteen percent of the gap). After one step the rate is 2.6%, after two 3.11%, after three about 3.5%; it keeps closing the gap but never overshoots. In regressions this is why last period's rate is such a strong predictor of this period's rate.

The Atlanta Fed Taylor Rule Utility

An online tool from the Federal Reserve Bank of Atlanta that computes the Taylor-rule rate for the US since 1985 and plots it next to the actual federal funds rate. You choose the inflation measure, target, r*, gap measure, weight and smoothing (up to three rules at once). Episodes to look at:

2003-2005: the actual rate was below the rule: rates too low before the financial crisis, Taylor's own criticism.
2009-2015: the rule prescribed negative rates while the Fed was stuck at zero (the ZLB).
2021-2022: the rule rose well before the actual rate: was the Fed "behind the curve" on inflation?

There is no such tool for the ECB; in the course you build that rule yourself.

The four questions of the course

Mandate: does the Fed react more to the output gap than the ECB, as its dual mandate suggests?
Exchange rate: does the exchange rate enter the rule, although neither mandate mentions it?
Symmetry: do central banks react equally strongly to inflation above and below target? Did this change after 2021 (ECB) and 2025 (Fed)?
Data or forecasts: do they react to current inflation or to their own forecasts?

And for the euro: one interest rate for twenty-plus countries in different situations. For whom is the ECB's rate "right"?

Andere hoofdstukken van International Monetary Relations and Policy

Test jezelf met oefenvragen