The authors regret the inclusion of a coding error in “The New Keynesian Wage Phillips Curve”. The composite inflation measure was correctly defined in the paper as

(5.6) \begin{align}\Pi_t^c \equiv \left(\Pi_t^p\right)^\vartheta\left(\Pi_t^w\right)^{1 - \vartheta} , \end{align}

where the weight $\vartheta \equiv \Lambda_p/(\Lambda_p + \Lambda_w)$ is given by the relative slopes of the linearized Price and Wage Phillips Curves, $\Lambda_p$ and $\Lambda_w$ , respectively. However, the computer code had switched the exponents and instead used

(5.6’) \begin{align} \Pi_t^c \equiv \left(\Pi_t^p\right)^{1-\vartheta}\left(\Pi_t^w\right)^{\vartheta} .\end{align}

Instead of putting an 85% weight on price inflation, only 15% was put on it. This mistake leaves the qualitative results of the paper unaffected. Quantitatively, the correction moves the numerical results for the composite case in Tables 4 to 6 closer to the inflation targeting case. The corrected tables are provided below.

Table 4.

Welfare: Efficient Steady State

Table 5.

Welfare: Inefficient Steady State

Table 6.

Model moments from the Gali (2015), Chapter 6 model

Notes: Displayed are the variance of log price inflation $\pi_{p}$ , log wage inflation $\pi_{w}$ , and of the log output gap $\tilde{y}$ . Numbers have been multiplied by 100.