Comment on ‘ Market discipline and monetary policy ’ by Carl Walsh : a reply

In their interesting paper, Fabia de Carvalho and Maurício Bugarin alter the infor-mational assumptions I employed in Walsh (2000) and discuss how this change affects the results reported in my paper. The structure of the underlying model I used, and therefore the rationale for the informational assumption, was designed to capture two factors. First, inflation models developed in the new Keynesian literature emphasize the role of expected future inflation and my model was designed to investigate the difference arising from using the forward-looking new Keynesian inflation models rather than the earlier literature's Lucas supply curve. In the new Keynesian approach, the informational assumption parallels the one I made in my paper— in setting prices (or wages) in period t, agents have information about period t variables. As the authors of this comment note, there is a resulting asymmetry in the informational structure, one that correctly captures the difference between the Lucas supply function and the new Keynesian inflation equation. Making the information assumptions the same for both types of wage models would have simply produced a model in which a simple Lucas supply function would be replaced by something more akin to an older-style multi-period wage or price contracting model. Second, with policy implemented via interest rates, a policy action is immediately observable in financial markets and leads to an immediate changes in the expectations embedded in interest rates. This immediate reaction cannot be captured with the assumption that only lagged information-sets matter. By altering the information assumptions, Carvalho and Bugarin employ a structure that is perhaps more intuitive—all contracts are based on past information rather than some on past information and some on contemporaneous information—but the resulting model misses out on one of the factors on which I focused—the notion that, when the central bank uses an interest rate to implement policy, markets can react immediately to policy changes and so they cannot capture the implications of moving from the Lucas supply function to the new Keynesian inflation equation. In contrast, the symmetric informational assumption employed by Carvalho and Bugarin shifts the focus onto the role of multi-period wage contracts directly.