Why Scenario Analysis Matters
A monetary-policy decision depends on more than whether inflation is above target. The cause matters. Strong demand can raise both output and inflation, while an adverse supply shock can raise inflation as output falls. A shift in inflation expectations can make an initially temporary price increase more persistent. The same policy-rate move therefore need not be appropriate in each case.
Scenario analysis makes those differences explicit. The policymaker identifies the shock, states the objective, estimates how available instruments affect the economy over the relevant horizon, and weighs inflation risks against output risks. A model can organize this reasoning, but its assumptions and uncertainty must remain visible. This tutorial is aligned with the SOA VEE Economics (Validation by Educational Experience) syllabus.
A Compact Model for Comparing Scenarios
Use the following deliberately simple, one-period model. All gaps and shocks are measured in percentage points. Let \(x_0\) be the initial output gap, \(D\) the demand shock, \(\Delta r\) the policy-rate change, \(e\) the contribution from a change in inflation expectations, and \(s\) the contribution from a supply shock. A positive \(\Delta r\) is a rate increase. For the examples, assume the policy rate affects the output gap with a lag represented by the coefficient \(b=0.5\), and assume the short-run inflation response to the output gap is \(\kappa=0.5\).
Here \(x_1\) is the output gap at the policy horizon, \(z_1\) is the inflation gap, \(\pi_1\) is inflation, and \(\pi^*\) is the inflation target. The model treats \(e\) and \(s\) as their contributions to inflation at that horizon, not as complete descriptions of expectations or supply conditions. The equations assume a linear response, a fixed target, and unchanged shocks under the candidate policy actions. They are a comparison tool, not a forecast or a structural law that applies in every economy.
To make the trade-off concrete, use the illustrative flexible-targeting loss \(L=z_1^2+x_1^2\). Lower loss is preferred. Giving the inflation gap and output gap equal squared weights is a simplifying assumption for these examples; an actual policymaker may use different weights and consider additional risks. The policy menu in each example is also explicitly limited to the listed rate changes.
A Consistent Method for Choosing a Response
Separate demand, supply, and expectations contributions. Check whether output and inflation are moving in ways consistent with one shock or a mixture.
Identify the inflation target, the output risk being considered, and the period over which the policy is expected to work. Monetary policy lags make timing part of the decision.
Trace each candidate action through the model. A rate change mainly affects demand in this setup; it does not directly remove a supply shock.
Explain which risks the decision addresses, what evidence could change it, and how the response would adjust if the shock proves more persistent than expected.
This sequence prevents the common error of selecting an instrument before diagnosing the shock. It also separates the objective from the instrument: a central bank can remain committed to its inflation objective while choosing a gradual response to limit unnecessary output volatility.
Why the Model Produces a Trade-Off
A short derivation shows how the loss criterion handles an inflation pressure that is not caused by demand. Let \(q=e+s\), so \(z_1=q+\kappa x_1\). Suppose, temporarily, that the policymaker can choose any output gap through the policy rate. Substituting the inflation equation into the loss gives \(L=(q+\kappa x_1)^2+x_1^2\). Differentiating with respect to \(x_1\), an interior minimum must satisfy:
Collecting the terms involving \(x_1\) gives \(2\kappa q+2(\kappa^2+1)x_1=0\). Therefore the minimizing output gap in this particular model is:
When \(q>0\), the model favors a negative output gap to reduce inflation pressure, but the squared output-gap cost prevents it from recommending that inflation be eliminated regardless of the effect on activity. The rate change required to achieve that gap is \(\Delta r=(x_0+D-x_1^*)/b\), if that action is available. In practice, the rate menu, uncertainty, model limitations, and changing shocks may make the calculated minimum unattainable or inappropriate.
Worked Example: Responding to a Positive Demand Shock
A hypothetical economy has an initial output gap \(x_0=0\), a demand shock \(D=2.0\), no supply or expectations contribution (\(s=e=0\)), and an inflation target of \(2.0\%\). Compare policy-rate increases of 0, 1, and 2 percentage points using the model and loss criterion above. Choose the action with the lowest loss from this menu.
Method. Calculate the output gap, inflation gap, and loss for each action. This method avoids assuming that a rate increase directly removes demand; in the model it reduces the output gap by \(0.5\) percentage point for each percentage point of rate increase.
With no increase, \(\Delta r=0\), so \(x_1=0+2.0-(0.5)(0)=2.0\). Then \(z_1=0+0+(0.5)(2.0)=1.0\), giving inflation of \(2.0\%+1.0\%=3.0\%\). The loss is \(L=(1.0)^2+(2.0)^2=1.0+4.0=5.0\).
With a 1-point increase, \(x_1=0+2.0-(0.5)(1)=1.5\). Then \(z_1=0+(0.5)(1.5)=0.75\), so inflation is \(2.75\%\). The loss is \(L=(0.75)^2+(1.5)^2=0.5625+2.25=2.8125\).
With a 2-point increase, \(x_1=0+2.0-(0.5)(2)=1.0\). Then \(z_1=0+(0.5)(1.0)=0.50\), so inflation is \(2.50\%\). The loss is \(L=(0.50)^2+(1.0)^2=0.25+1.00=1.25\).
The lowest loss among the available choices is \(1.25\), so the model selects a 2-percentage-point increase. It leaves both inflation and output above their reference levels, rather than eliminating the shock entirely. This menu-based result does not establish that such a large move is feasible or prudent in an actual economy; it depends on the stipulated horizon and coefficients.
Worked Example: Weighing a Temporary Supply Shock
Suppose a hypothetical disruption to imported components contributes \(s=1.5\) percentage points to inflation at the policy horizon. The initial output gap and demand shock are both zero, and \(e=0\). The target is \(2.0\%\). Compare rate increases of 0, 1, and 2 percentage points. Assume the supply contribution remains 1.5 points over this horizon, even though policymakers believe it may later fade.
Method. Apply the same loss to each candidate action, but keep the supply contribution separate from the output channel. This avoids treating a rate increase as if it could restore the disrupted supply of components.
With \(\Delta r=0\), \(x_1=0+0-(0.5)(0)=0\). Inflation gap is \(z_1=1.5+(0.5)(0)=1.5\), so inflation is \(3.5\%\). The loss is \(L=(1.5)^2+0^2=2.25\).
With \(\Delta r=1\), \(x_1=0+0-(0.5)(1)=-0.5\). The inflation gap is \(z_1=1.5+(0.5)(-0.5)=1.25\), so inflation is \(3.25\%\). The loss is \(L=(1.25)^2+(-0.5)^2=1.5625+0.25=1.8125\).
With \(\Delta r=2\), \(x_1=0+0-(0.5)(2)=-1.0\). The inflation gap is \(z_1=1.5+(0.5)(-1.0)=1.0\), so inflation is \(3.0\%\). The loss is \(L=(1.0)^2+(-1.0)^2=1.00+1.00=2.00\).
The 1-point increase has the lowest loss of the three options. The additional increase to 2 points lowers the inflation gap from 1.25 to 1.0, but pushes the output gap from \(-0.5\) to \(-1.0\), enough to raise the stipulated loss. A careful communication strategy would explain the response to inflation persistence and expectations, while distinguishing those concerns from the temporary supply disruption itself.
Worked Example: Responding to an Expectations Shock
A hypothetical economy begins with \(x_0=0\), \(D=0\), and \(s=0\). New information suggests that the expectations contribution to inflation at the policy horizon has risen to \(e=1.2\) percentage points. The inflation target is \(2.0\%\). Compare rate increases of 0, 1, and 2 percentage points, and identify a communication approach consistent with the numerical result.
Method. Calculate the model outcome and loss under each action, then communicate both the inflation risk and the output cost. This avoids responding to an expectations shift with reassurance alone when the model indicates that policy action can also matter.
With \(\Delta r=0\), \(x_1=0\), \(z_1=1.2+(0.5)(0)=1.2\), and inflation is \(3.2\%\). The loss is \(L=(1.2)^2+0^2=1.44\).
With \(\Delta r=1\), \(x_1=0-(0.5)(1)=-0.5\). Then \(z_1=1.2+(0.5)(-0.5)=1.2-0.25=0.95\), so inflation is \(2.95\%\). The loss is \(L=(0.95)^2+(-0.5)^2=0.9025+0.25=1.1525\).
With \(\Delta r=2\), \(x_1=-(0.5)(2)=-1.0\). Then \(z_1=1.2+(0.5)(-1.0)=0.70\), so inflation is \(2.70\%\). The loss is \(L=(0.70)^2+(-1.0)^2=0.49+1.00=1.49\).
The model selects a 1-point increase because \(1.1525\) is the lowest loss. Communication should describe the concern that higher expected inflation could persist, explain why a measured increase is preferred to the larger option under the stated trade-off, and say what evidence about expectations and inflation persistence would prompt a revision. A conditional explanation is more informative than a promise to follow a fixed rate path regardless of new data.
Match Instruments and Communication to the Shock
The policy rate is not the only available tool, and not every tool solves the same problem. If the policy rate is near its effective lower bound, earlier material on the effective lower bound and large-scale asset purchases is relevant to the choice of instruments. Forward guidance can affect expectations of future policy rates, while asset purchases can influence longer-term financial conditions. These instruments may support demand; none should be described as directly repairing a physical supply disruption.
Communication should identify the diagnosis without overstating certainty. A demand-driven inflation increase may justify a clear explanation of why restraint is needed. A supply shock calls for explaining the tension between inflation and output, and for stating what evidence would make a persistent response necessary. An expectations shock makes the policy commitment and the conditions for future action especially important. In each case, the central bank should distinguish its objective from its uncertain estimate of the economy's response to an instrument.
| Scenario | Primary risk | Policy reasoning | Communication emphasis |
|---|---|---|---|
| Positive demand shock | Excess demand raises output and inflation | Restrict demand if inflation risk outweighs the cost to activity | Explain the intended demand response and policy horizon |
| Adverse supply shock | Inflation rises while output may weaken | Weigh shock persistence and expectations against the cost of further restraint | Explain the trade-off and the conditions that change the response |
| Expectations shock | Inflation pressure may persist through expectations | Assess whether action is needed to keep expectations consistent with the objective | State the commitment, uncertainty, and data that would prompt revision |
Common Traps in Scenario Analysis
- Using one response for every inflation increase. Inflation above target does not identify its cause. Diagnose the demand, supply, and expectations components before choosing an action.
- Confusing a shock with a policy instrument. In the model, the rate changes demand and output; it does not directly change \(s\). Claiming that a rate increase fixes the supply disruption confuses the inflation channel with the source of the shock.
- Ignoring output costs. A policy that lowers the inflation gap can still have a higher total loss if it creates a sufficiently negative output gap.
- Presenting a stylized calculation as a precise forecast. The coefficients, horizon, shock estimates, and loss weights are assumptions. A real decision should consider uncertainty and alternative scenarios.
- Giving unconditional guidance when the diagnosis is uncertain. A credible explanation states what is known, what remains uncertain, and which changes in evidence would alter the response.
Check Your Understanding
Use \(x_1=x_0+D-0.5\Delta r\), \(z_1=e+s+0.5x_1\), inflation target \(2.0\%\), and \(L=z_1^2+x_1^2\), unless a question states otherwise. All figures are percentage points.
- An economy has \(x_0=-1.0\), \(D=1.5\), and \(e=s=0\). If the central bank raises the rate by \(0.5\) point, calculate the output gap, inflation gap, and inflation. (Answer: \(x_1=0.25\), \(z_1=0.125\), inflation \(=2.125\%\).)
- A supply shock contributes \(s=2.0\), with \(x_0=D=e=0\). Compare rate increases of 0, 1, and 2 points using the stated loss; which is preferred? (Answer: losses are \(4.00\), \(3.3125\), and \(3.25\), respectively; choose a 2-point increase from this menu.)
- A positive demand shock is \(D=1.2\), with \(x_0=0\), \(e=0.2\), and \(s=0\). For rate increases of 0, 1, and 2 points, calculate the loss and select the lowest-loss action. (Answer: losses are \(2.08\), \(0.7925\), and \(0.13\), respectively; choose a 2-point increase.)
- There is an expectations contribution \(e=0.8\) and supply contribution \(s=0.4\), with \(x_0=D=0\). Compare rate increases of 0, 1, and 2 points. Which action minimizes the loss, and what should communication explain? (Answer: losses are \(1.44\), \(1.1525\), and \(1.49\); choose a 1-point increase and explain both the combined inflation risk and the output cost.)
- Let \(x_0=0\), \(D=1.0\), \(e=0\), and \(s=1.0\). Using the derived unconstrained minimizing output gap, calculate the rate increase that would attain it under the model. If the feasible menu is limited to increases no greater than 2 points, can the model's unconstrained minimum be attained? (Answer: \(x_1^*=-0.4\), the required increase is \(2.8\) points, and it cannot be attained with the stated menu.)