Comments and Discussion

This is a paper on the way data relate to decisionmaking in central banks. One component of the paper is based on a series of interviews with staff members and a few policy committee members of four central banks: the Swedish Riksbank, the European Central Bank (ECB), the Bank of England, and the U.S. Federal Reserve. These interviews focused on the policy process and sought to determine how forecasts were made, how uncertainty was characterized and handled, and what role formal economic models played in the process at each central bank.

In each of the four central banks, "subjective" forecasting, based on data analysis by sectoral "experts," plays an important role. At the Federal Reserve, a seventeen-year record of model-based forecasts can be compared with a longer record of subjective forecasts, and a second component of this paper is an analysis of these records.

Two of the central banks—the Riksbank and the Bank of England—have explicit inflation-targeting policies that require them to set quantitative targets for inflation and to publish, several times a year, their forecasts of inflation. A third component of the paper discusses the effects of such a policy regime on the policy process and on the role of models within it.

The large models in use in central banks today grew out of a first generation of large models that were thought to be founded on the statistical theory of simultaneous-equations models. Today's large models have completely lost their connection to this theory, or indeed to any other probability-based theory of inference. The models are now fit to data by ad hoc procedures that have no grounding in statistical theory. A fourth component of the paper discusses how inference using these models reached this state and why academic econometrics has had so little impact in correcting it. Despite their failure to provide better forecasts, their lack of a firm statistical foundation, and the weaknesses in their underlying economic theory, the large models play an important role in the policy process. A final component of the paper discusses what this role is and how the model's performance in it might be improved.

The Policy Process

At all four central banks the policy process runs in a regular cycle; that cycle is quarterly in frequency at all except the Federal Reserve, where it is keyed to the meetings of the Federal Open Market Committee (FOMC), which take place roughly every six weeks. Each central bank has a primary macroeconomic model but uses other models as well. The primary models are the ones used to construct projections of alternative scenarios, conditional on various assumptions about future disturbances or policies or on various assumptions about the current state of the economy. Where there is feedback between models and subjective forecasts, it is generally through the primary model.

The primary models have some strong similarities. The ECB's model contains about fifteen behavioral equations, the Bank of England's twenty-one, the Riksbank's twenty-seven, and the Federal Reserve's about forty.¹ Each has at least some expectational components, with the Federal Reserve and Riksbank models the most complete in this respect. Those central banks whose models are less forward-looking describe them somewhat apologetically, suggesting that they are working on including more forward-looking behavior.

The Riksbank and the Bank of England have publicly described "suites" of models of various types, including vector autoregressive (VAR) models, smaller macroeconomic models, and optimizing models. Some of these models produce regular forecasts that are seen by those involved in the policy process, but at both central banks none except the primary model has a regular, well-defined role. The other central banks also have secondary models with some informal impact on the policy process.

Each policy round proceeds through a number of meetings, through which a forecast is arrived at iteratively, but the number of meetings and the way discussions are ordered vary. At the Riksbank there is a start-up meeting, at which forecasts from two large models are presented, followed...