Why Stabilization Requires a Counterfactual
Output and inflation data describe what happened, but they do not by themselves show what fiscal policy changed. If output falls less than expected after a spending program, the difference could reflect the program, a smaller shock, a change in monetary policy, or some combination. To assess stabilization, compare the observed outcome with a clearly defined estimate of what would have happened without the fiscal action.
The comparison must respect timing. A measure cannot affect output before it takes effect, and an announced measure may take time to be implemented or to influence spending. The comparison must also specify how monetary policy is treated: a central bank may respond to the fiscal action, so the outcome with that response can differ from the outcome with the monetary setting held fixed.
Define the Outcome and the Comparison
Let \(Y_t\) denote real output in period \(t\), and let \(Y_t^*\) denote potential output, the level of output consistent with the economy’s productive capacity under the measure being used. The output gap, expressed as a percentage of potential output, is
A negative gap means output is below potential; a positive gap means it is above potential. Superscripts \(0\) and \(1\) distinguish the no-policy counterfactual and the policy outcome. For example, \(g_t^0=-5\%\) and \(g_t^1=-3\%\) mean the output gap is two percentage points less negative with the policy. This is a change in percentage points, not a 2% change in output.
A convenient measure of stabilization in a single period is the proportional reduction in the absolute output gap:
A positive \(R_t\) indicates a smaller gap in absolute value; a negative value indicates that the policy outcome is farther from potential. The percentage is undefined when the counterfactual gap is zero. If the policy moves the economy across zero, report both gaps and the absolute-gap change rather than relying on a percentage alone: a smaller signed gap is not necessarily a smaller absolute gap.
Deriving the Gap-Reduction Measure
Suppose the no-policy gap is negative, \(g^0=-L\), where \(L>0\). Suppose the fiscal action raises the gap by \(d\) percentage points, so \(g^1=-L+d\). If \(0\le d\le L\), the policy outcome remains at or below potential. Its absolute gap is \(L-d\). Therefore, the decrease in the absolute gap is \(L-(L-d)=d\), and the proportional decrease is
This derivation explains why a one-percentage-point improvement means different proportional reductions from different starting gaps: it is a 20% reduction from a five-point gap, but a 10% reduction from a ten-point gap. The result assumes the gap does not cross zero. If \(d>L\), then \(g^1\) is positive and \(|g^1|\) is no longer \(L-d\); calculate both absolute gaps directly.
Work Through Timing Before Attribution
A counterfactual comparison needs a time path, not just a before-and-after observation. Recognition, decision, implementation, and impact lags can mean that a policy has no effect in one period but does affect a later period. A difference that appears before the stated impact date must be explained by another factor or by a revised timing assumption; it cannot be credited to a measure that has not yet affected output.
The following worked example treats the counterfactual as an estimate under the same background shocks and monetary-policy rule. Thus, a monetary response to the fiscal action is included in the policy path if the evaluation is of the action’s total effect under that rule.
Worked Example: Evaluating a Program With an Impact Lag
A temporary public-maintenance program is implemented during a downturn. Its estimated output effect begins in quarter 3, not before. The no-program counterfactual output gaps are \(-8\%\), \(-6\%\), and \(-4\%\) in quarters 1, 2, and 3. The observed gaps with the program are \(-8\%\), \(-6\%\), and \(-3\%\). Assume the counterfactual and observed paths face the same shocks and monetary-policy rule. Find the quarter 3 reduction in the absolute gap and the cumulative reduction across the three quarters.
Method. Compare the two paths period by period, first checking that their differences are consistent with the program’s stated impact timing.
Timing check. In quarter 1, both gaps are \(-8\%\); in quarter 2, both are \(-6\%\). There is no difference before the program’s estimated impact begins. In quarter 3, the counterfactual gap is \(-4\%\) and the observed gap is \(-3\%\), a one-percentage-point improvement.
Quarter 3 calculation. The absolute counterfactual gap is \(4\) percentage points and the absolute policy gap is \(3\) percentage points. The absolute-gap reduction is \(4-3=1\) percentage point, so
Cumulative calculation. Using the sum of absolute gaps as the measure over these three equally weighted quarters, the counterfactual total is \(8+6+4=18\) percentage-point-quarters. The policy total is \(8+6+3=17\) percentage-point-quarters. The cumulative reduction is \(18-17=1\), or
The program is associated with a 25% reduction in the quarter 3 gap and a 5.56% reduction in the cumulative absolute gap over the stated horizon, conditional on the counterfactual assumptions. The smaller cumulative percentage reflects the two earlier quarters with no difference. Common mistake avoided: attributing an earlier-period difference to a measure whose output effect has not yet begun. Here the pre-impact gaps match, as the timing assumption requires.
Separate Fiscal Effects From Monetary Responses
There are two useful questions, and they are not interchangeable. A comparison that holds the monetary setting fixed estimates the fiscal action’s effect without a monetary response. A comparison that allows monetary policy to respond estimates the total effect under that response rule. For the second question, do not treat the observed output change as the fiscal effect alone: distinguish the fiscal contribution from the monetary contribution when those estimates are available.
The comparison should state which question it answers. If the central bank reacts differently in the policy and no-policy paths, holding its response fixed in one path while allowing it to change in the other would not be a consistent comparison.
Worked Example: Allowing for a Monetary Offset
In a hypothetical economy, the no-fiscal-action counterfactual has an output gap of \(-5\%\) this year. A temporary transfer program is estimated to raise the gap by \(2\) percentage points if the monetary setting is held fixed. Under the stated monetary-policy rule, the central bank’s response lowers the gap by \(0.8\) percentage points relative to that fixed-setting outcome. Find the policy outcome and its gap reduction relative to the no-action counterfactual.
Method. Add the fiscal contribution and the monetary response to the same no-policy starting gap. The monetary response is negative here because it offsets part of the fiscal expansion.
Fixed-monetary-setting outcome. Starting at \(-5\%\), the fiscal contribution of \(+2\) percentage points gives \(-5+2=-3\%\).
Outcome with the monetary response. The response lowers the gap by another \(0.8\) percentage points: \(-3-0.8=-3.8\%\). Equivalently, the net change from the no-policy counterfactual is \(+2-0.8=+1.2\) percentage points, and \(-5+1.2=-3.8\%\).
Stabilization calculation. The absolute gap falls from \(5\) to \(3.8\) percentage points, a reduction of \(1.2\) points. Thus,
The monetary response offsets \(0.8/2=40\%\) of the estimated \(2\)-point fiscal contribution. The total gap reduction under the stated monetary rule is 24%, not the 40% that would result from incorrectly using the fixed-setting gap of \(-3\%\) as the final policy outcome. Common mistake avoided: attributing the entire fixed-monetary-setting effect to fiscal policy when the question asks for the outcome after the monetary response.
Positive Gaps and Multi-Period Results
Stabilization also applies when output is above potential. In that case, a fiscal action that reduces a positive gap can moderate an expansion. The same absolute-gap measure works: compare distances from zero, rather than assuming that every beneficial policy must raise output.
For multiple periods, one simple summary is the sum of absolute gaps, as in the first example. If periods have different lengths or importance, a weighted average may be more appropriate, but the weights should be specified. A cumulative statistic can conceal an increase in the gap in one period, so inspect the period-by-period results as well as the summary.
Worked Example: Moderating an Above-Potential Expansion
A hypothetical economy has a no-action output gap of \(+4\%\). A fiscal restraint is estimated to reduce the gap by \(2.5\) percentage points. In response, monetary policy accommodates part of the restraint, adding \(0.5\) percentage points back to the gap. Calculate the final gap, the proportionate reduction in its absolute size, and the portion of the fiscal contribution offset by the monetary response.
Method. Start with the no-action counterfactual and combine the signed contributions. Fiscal restraint contributes \(-2.5\) points; monetary accommodation contributes \(+0.5\) points.
Final gap. The net change is \(-2.5+0.5=-2.0\) percentage points. Therefore, the policy outcome is \(+4-2=+2\%\), still above potential.
Gap reduction. The absolute gap falls from \(4\) to \(2\) percentage points. The reduction is \(4-2=2\) points, so
Monetary offset. The response offsets \(0.5/2.5=0.20\), or 20%, of the fiscal contribution. The resulting 50% reduction measures the change in the gap after that response, relative to the no-action counterfactual. Common mistake avoided: concluding that a fiscal action cannot be stabilizing because it lowers output. Here output starts above potential, so reducing the positive gap is the relevant direction.
When a Difference Is Not Yet a Policy Effect
A measured difference between the observed and no-policy paths is informative only if the comparison is credible. The counterfactual is not directly observed; it must be estimated, for example from a model, a comparison group, or a forecast. Each approach depends on assumptions. If the no-policy forecast assumes a larger shock than the one that occurred, its gap may be too negative, making the fiscal action appear more effective than it was.
Check the following before interpreting a calculated reduction:
- Timing: Does the estimated difference begin only when the policy could plausibly affect output?
- Common conditions: Do the two paths use consistent assumptions about external shocks, potential output, and other policies?
- Monetary response: Is the comparison holding monetary policy fixed, or allowing a specified response in both paths?
- Metric: Are you comparing signed gaps, absolute gaps, or a cumulative measure? State which one.
- Uncertainty: Is the counterfactual an estimate with uncertainty? A precise arithmetic result does not make uncertain causal assumptions certain.
The calculation should be interpreted as conditional on the counterfactual and its assumptions. It does not prove that all of the observed difference was caused by fiscal action. This distinction matters when evaluating policy effectiveness: a smaller output gap is a favorable outcome, but attribution requires more than observing that it occurred after a fiscal measure.
Check Your Understanding
For each problem, state the relevant comparison before interpreting the arithmetic.
- A training subsidy is estimated to have no output effect in quarter 1 and to affect output in quarter 2. The no-subsidy gaps are \(-5\%\) and \(-4\%\); the observed gaps are \(-5\%\) and \(-3\%\). What is the quarter 2 proportional reduction in the absolute gap, and are the paths consistent with the stated timing? (Answer: 25%; yes, the paths match in quarter 1 and first differ in quarter 2.)
- A no-action counterfactual has a gap of \(-6\%\). A fiscal measure raises the gap by \(1.5\) percentage points, while the monetary response lowers it by \(0.3\) points. Find the final gap and the absolute-gap reduction as a percentage of the counterfactual gap. (Answer: \(-4.8\%\); 20%.)
- The no-policy gaps over three equally weighted quarters are \(-4\%\), \(-6\%\), and \(-5\%\). The policy gaps are \(-4\%\), \(-6\%\), and \(-3.5\%\). Find the cumulative proportional reduction in the sum of absolute gaps. (Answer: \(1.5/15=10\%\).)
- An economy’s no-action gap is \(+3\%\). Fiscal restraint lowers it by \(1.8\) percentage points; a monetary response raises it by \(0.3\) points. Find the final gap, the absolute-gap reduction, and the proportion of the fiscal contribution offset. (Answer: \(+1.5\%\); 50%; \(0.3/1.8=16.67\%\), or approximately 16.7%.)
- Across two periods, the no-policy gaps are \(-2\%\) and \(-6\%\), while the policy gaps are \(-2\%\) and \(-5\%\). Using equal period weights, calculate the cumulative proportional reduction in the sum of absolute gaps, and identify whether each period improved. (Answer: \((8-7)/8=12.5\%\); period 1 is unchanged and period 2 improves.)