Choosing Which Future to Show
The Federal Reserve's alternative scenarios are not neutral windows onto every possible future; they are choices about which uncertainty to make visible and how much detail to attach to it. This chapter focuses on the criteria behind that choice—risk identity, mechanism, quantifiability, severity, information value, and likely misunderstanding. The reader should come away able to judge when a fan chart or qualitative discussion is enough, when an alternative scenario adds necessary structure, and why a scenario close to the baseline solves a different problem from an intentionally severe departure.
Match the tool to the uncertainty
A central bank starts with a practical constraint: no single communication device can express every kind of uncertainty equally well. The Federal Reserve's historical practice places alternative scenarios beside a baseline, while the BIS comparison shows how fan charts, qualitative risk discussions, and scenario analysis occupy different roles across central banks. The choice is not between being quantitative and being vague. It is between different kinds of information and different ways that information can fail.
A fan chart is the natural first choice when the question is broad forecast dispersion. It puts the central outlook inside intervals usually informed by past forecast errors and signals that outcomes can vary around the central line. This is useful when the main message is that the baseline is uncertain in the ordinary sense. It is concise, comparable across forecast releases, and less likely than a long list of branches to imply that the institution has selected one particular story about the future.
The fan chart is less suitable when the important uncertainty has separate, recognizable branches. A central band may average together a timely and a delayed vaccine, or temporary and persistent inflation, even though those paths would have different consequences for policy. It also cannot easily show why an inflation outcome is associated with stronger output in one case and weaker output in another. The chart communicates the range but can hide the cause.
An alternative scenario is a better fit when the cause is specific enough to state. It can take a tariff change, a productivity slowdown, a change in household saving, or a structural shift in price-setting and carry that assumption through an economic model. The result is a path for output, inflation, and possibly other variables relative to the baseline. The reader can inspect the transmission rather than merely observe a wider interval.
That advantage depends on the risk being identifiable and the mechanism being sufficiently credible to quantify. A scenario that assigns precise numbers to a development whose timing, size, or transmission is unknowable may create false confidence. The problem is not that the model has uncertainty; the problem is that the display can make the uncertainty disappear behind precise figures. In such a case, a qualitative risk discussion may be the better fit.
A qualitative discussion can name a trigger, explain plausible channels, and describe whether a risk points toward worse or better outcomes without claiming that its magnitude has been measured. It is particularly useful for novel, politically contingent, or structurally ambiguous developments.