The Baseline, the Branches, and the Toolkit

The Federal Reserve's alternative-scenario practice becomes easier to understand once the different kinds of uncertainty are separated. This chapter focuses on the mechanism by which a central bank matches a communication tool to a question: is the uncertainty a broad spread around the outlook, a named risk with a distinct transmission path, or doubt about how policy will respond? The distinction gives readers a way to tell what a fan chart, a qualitative risk discussion, an alternative scenario, or a policy path can legitimately show.

Two kinds of uncertainty, not one

No economic forecast can be exact because the future is inherently unpredictable. Even if a central bank understood its models and current data perfectly, households, firms, governments, and other economies would not follow a script. The economy will rarely evolve exactly along the baseline. This is general uncertainty: the unavoidable possibility that the central path will be too high, too low, too early, or too late.

A second kind is specific uncertainty. It comes from an identifiable development that could alter the economy and the monetary-policy stance: a tariff decision, a sudden increase in energy prices, a financial-market correction, or a structural change in how prices are set. The event may itself be uncertain, but it has an identity. Policymakers can ask what it would mean if that particular development occurred.

The distinction is practical. General uncertainty calls for a description of the range around a forecast. Specific uncertainty calls for an explanation of the branch that creates the range. A central bank that treats both as the same problem risks either flattening a meaningful story into a statistical band or turning every ordinary forecast error into a dramatic scenario.

A simple way to see the division is to ask two questions. First, “How widely could outcomes vary around the central outlook?” Second, “What particular mechanism could move outcomes away from it?” A fan chart is designed mainly for the first question. An alternative scenario is designed mainly for the second. A qualitative discussion may be the most honest answer when the mechanism is identifiable but cannot be quantified credibly.

This distinction also prevents a common reading error: confusing a display of uncertainty with a set of probabilities. A range generated from past forecast errors may describe how far a forecast has tended to miss, but it does not necessarily say that each point inside the range is equally plausible or that every possible cause has been represented. Likewise, publishing a tariff, supply, or financial scenario shows that the institution considers the branch decision-relevant; it does not establish that the event is expected or that the branch exhausts the relevant risks.

For the Federal Reserve, the useful question is therefore not whether an alternative is “the forecast.” It is what assumption has been changed and what the comparison teaches about the baseline. If the assumption is a shock, the scenario can expose the economy's sensitivity to that shock. If the assumption is a relationship—how expectations form, how prices adjust, or how financial conditions transmit—it can expose model uncertainty. In both cases, the changed assumption is the hinge of the analysis.

What each outlook tool can reveal

A fan chart places a central forecast inside bands that represent uncertainty around it. The intervals are typically estimated from past forecast errors. That makes the display compact and useful for communicating that a forecast is not a point of certainty. It can help the public understand that an inflation or growth outcome outside the central line is not automatically evidence of analytical failure; some variation is built into forecasting.

But the same compression is the fan chart's limitation. A band tells the reader how far outcomes might spread, not necessarily why. If a chart is centered on a baseline that averages across distinct possible futures, it may conceal the fact that the economy could follow two very different paths. A timely vaccine and a delayed vaccine are not simply two random draws from one smooth distribution. Nor are temporary inflation, persistent inflation, and a wage–price spiral the same risk with different sizes. They may require different reasoning about policy even if their average is represented by one central line.

Fan charts can also become difficult to interpret after unusually large shocks. When historical forecast errors are used to set the width, including extreme observations can make the bands much wider. That may be statistically defensible as a record of past errors, yet the visual result can dilute the central bank's message about the outlook and raise questions about its forecasting ability. Adjusting for exceptional shocks can narrow the display, but then the institution must explain why those observations should not define the ordinary range.

An alternative scenario preserves the source of the uncertainty. It starts with a changed assumption—a rise in tariffs, slower productivity, weaker household spending, or a shift in financial risk premiums—and follows the implications through a model or a model extension. The result is another structured point forecast, not a second claim about what is most likely. Its purpose is to make a selected causal path concrete enough to compare with the baseline.

The construction can be read as a chain of conditional questions. What is different at the starting point? Which households, firms, markets, or foreign economies respond first? Through which prices, incomes, credit conditions, or expectations does the disturbance spread? Over what horizon do those effects appear? Finally, is the policy rate held on the baseline path, or is a response included? Each answer changes the meaning of the resulting numbers. A scenario that holds policy fixed isolates the economic cost of the disturbance under a common policy assumption; one with a response illustrates a different question about stabilization.

This is also why a scenario cannot be judged by its final GDP or inflation number alone. Two paths may end at a similar inflation rate while reaching it through different combinations of demand, supply, expectations, and policy. The intermediate mechanism tells policymakers whether the path represents a temporary deviation, a persistent problem, or a change in the relationships on which the baseline depends. The numbers are useful because they are attached to a story that can be questioned.

That comparison matters because macroeconomic variables do not move independently. Suppose households save less than expected. The resulting demand shock can raise both output and inflation. Suppose instead productivity grows more slowly. The supply shock can raise inflation while weakening output. Looking only at the inflation increase would miss the policy-relevant difference. The scenario tells the reader which economic force produced the movement and what else moved with it.

A scenario can also represent structural uncertainty rather than a single event. The question may be whether a familiar relationship in the model is still reliable—for example, whether firms adjust prices more frequently, making inflation more responsive to economic slack. Applying the same demand disturbance under a different structural relationship can produce a different inflation path. Here the alternative is not simply “a larger shock.” It is a conditional test of how the economy works.

A qualitative risk discussion occupies a different place. It can name the trigger, describe the channels through which it might matter, and indicate whether the risk appears to be on the upside or downside without claiming a numerical precision the evidence cannot support. That is valuable when the development is novel, politically contingent, or difficult to calibrate. Its weakness is the inverse of the scenario's strength: without a quantitative path, readers may struggle to judge the possible scale or compare the risk with the baseline.

The qualitative assessment can also carry information that a single scenario might overstate. It can acknowledge several possible triggers, competing channels, or uncertainty about magnitude without forcing them into one numerical path. A selected scenario can then show one concrete reason the outlook might depart from the baseline, while the accompanying discussion makes clear that the branch is not the whole distribution.

The three outlook tools therefore answer related but nonidentical questions:

This is not a ranking. The Federal Reserve's historical practice used quantitative alternatives alongside qualitative assessments, and the broader central-bank evidence points to the same complementarity. The tool is useful when it exposes information that another tool is structurally bad at carrying.

The choice of tool also affects how a reader interprets institutional confidence. A narrow-looking fan chart can be mistaken for certainty; a very wide one can look like an admission that the central bank has no usable view. A named scenario can make an otherwise abstract risk memorable, but its visibility may be mistaken for a claim that the risk is most likely. A qualitative warning can avoid false precision, but if it names no mechanism or consequence it becomes difficult to act on. Good communication keeps the tool's purpose and its limits adjacent to the display.

The same discipline applies to the baseline itself. It should remain the reference point, because alternatives have meaning only as departures from something. But the comparison should not imply that the baseline is a neutral fact while the scenario is a subjective story. Both depend on assumptions, models, and judgment. The scenario simply makes one altered assumption explicit, allowing readers and policymakers to ask whether that assumption is plausible and whether the modeled response is credible.

Policy uncertainty is a separate layer

The public does not need to understand only what might happen to growth and inflation. They also need to understand how the central bank might react. That creates a second uncertainty problem. Communication about the economic outlook tries to convey how the forecast could be wrong. Communication about the policy outlook often tries to reduce uncertainty about the reaction function—the relationship between economic developments and the decisions that follow.

Those goals can pull in opposite directions. More detail about the economic outlook may make the policy response seem easier to infer, but a central bank may still need room to respond to developments that the published outlook did not anticipate. Conversely, emphasizing flexibility can preserve room to maneuver while leaving households, businesses, and markets unsure how the institution will respond. The communication device is part of the policy transmission process because expectations about future rates affect financial conditions today.

Descriptive guidance addresses that second problem with words. It may indicate the likely direction of the policy stance or, in stronger forms, signal a commitment to keep rates on a path while certain conditions hold. Such guidance can be useful when the central bank wants to influence expectations about future rates. It can also be heard as an unconditional promise even when the institution intends it to be conditional.

That difference between an assessment and a commitment is important. An assessment says what policymakers currently expect to do if their outlook remains broadly valid. A commitment says they will follow a path, subject to whatever conditions have been stated. In a world of specific shocks, the conditions can change faster than ordinary language makes clear. The public may remember the path and forget the condition, turning a conditional statement into a test of credibility when the central bank changes course.

A central bank's own policy-rate projection uses numbers instead. It quantifies an expected path and links that path to the central forecast for inflation and real activity. In principle, this makes the relationship between the outlook and policy more visible and testable. It is different from a technical assumption that rates will remain constant or from a path taken from market pricing: the relevant projection is the central bank's own indication of how policy might evolve.

A scenario with a policy response is more specific still. It asks: if this particular shock occurs, how might policy react, and what would that reaction do to the economy? That can clarify conditionality. The public can see that the policy rate is not a free-standing promise; it depends on the development that actually occurs. It may also clarify the central bank's willingness to act when necessary to meet its objectives.

Yet the extra information creates extra failure paths. A policy response can change inflation and output, which then changes the policy response again. The displayed path is no longer just the effect of a shock. It is the result of a feedback loop among the shock, the economic model, the policy rule or judgment, and the variables that policy is trying to stabilize. A simple rule may be transparent while leaving out financial conditions or other considerations that matter in practice.

That is why none of these tools can carry the whole future. A fan chart can show broad dispersion but not a causal story. A qualitative discussion can preserve a novel risk but leave its magnitude open. An alternative scenario can make a mechanism and trade-off visible but gives prominence to a selected branch and does not tell readers how likely it is. A policy projection can clarify a central path but invite excessive attention to a number that the central bank may later need to change. A policy-response scenario can clarify conditionality but depends on assumptions about how policy would behave in an unfamiliar situation.

The most accurate reading of the Federal Reserve's alternatives is therefore not that the institution has found a better single forecast. It has added a way to keep specific mechanisms visible beside the forecast. The difference between a broad band and an explicit branch is the difference between knowing that uncertainty exists and understanding one important way it could unfold.

That distinction sets up the Federal Reserve's history. The next issue is not simply when more scenarios appeared, but how changes in models, reporting practice, and recurring risk assessment made the “what if?” question more usable inside the institution.

References

• Alternative Scenarios at the Federal Reserve from 1968 to 2020: Data, Interpretation, and Evaluation (PDF)
• Evolving approaches to monetary policy communication in the face of uncertainty: fan charts, scenarios and guidance