Why the Effective Lower Bound Changes Fiscal Stabilization
When demand is weak, a central bank can often support spending by reducing its policy interest rate. But there is a limit to how far the nominal rate can be cut: at the effective lower bound (ELB), further reductions are unavailable or judged impractical. A fiscal expansion may then raise output more than it would if the central bank were free to increase rates in response to stronger demand.
This is a conditional result, not a rule that every fiscal measure has a large effect at the lower bound. The size of the response depends on how much spending feeds into further demand, whether interest rates rise, and how households and firms respond. The analysis here uses a simple short-run model to make those channels explicit. It is aligned with the SOA VEE Economics (Validation by Educational Experience) syllabus.
A Short-Run Demand Model
Let \(x\) denote the change in the output gap, measured in percentage points of potential output. Let \(g\) denote the direct fiscal demand impulse, also measured in percentage points of potential output. For example, a government-purchases increase may create a positive \(g\). Let \(c\) be the share of an increase in output that feeds back into additional demand, and let \(b\) measure how strongly demand responds to a change in the real interest rate.
The reduced-form model is:
Here, \(\Delta r\) is the change in the real interest rate, measured in percentage points. In this model, a real-rate increase reduces demand, which explains the minus sign. Assume \(0 \leq c < 1\) and \(b \geq 0\). The model holds other influences on demand fixed and describes a short-run response; it does not capture every feature of an economy. In particular, its numerical multipliers are illustrative, not universal estimates.
For the basic comparison, assume expected inflation is fixed. Then a change in the nominal policy rate produces the same-sized change in the real rate. At a binding lower bound, assume the policy rate is held at that bound, so there is no policy-rate increase in response to the fiscal expansion. If the central bank instead raises rates as demand strengthens, the interest-rate increase partly offsets the fiscal impulse.
Why the Multiplier Is Larger When Rates Do Not Rise
At a binding lower bound, with the policy rate fixed and expected inflation unchanged, \(\Delta r=0\). Substituting this condition into the demand model gives the lower-bound fiscal multiplier. The multiplier is the change in the output gap for a one-unit change in the fiscal impulse.
The denominator \(1-c\) appears because an initial demand increase raises output, and the resulting income supports further spending. That spending raises output again, creating another round of demand. With \(c<1\), each round is smaller than the previous one, so the rounds form a convergent geometric series:
This also explains why the formula requires \(c<1\): if each round generated at least as much additional demand as the previous round, the series would not converge. The multiplier is larger when \(c\) is closer to one, but it is not infinite in this model.
Now suppose the central bank follows a simple response rule: it raises the nominal interest rate by \(\phi x\) percentage points when the output gap rises by \(x\) percentage points, where \(\phi \geq 0\). With expected inflation fixed, \(\Delta r=\phi x\). Substitution gives:
For \(b\phi>0\), this multiplier is smaller than \(1/(1-c)\). The reason is the interest-rate channel: stronger demand prompts a rate increase, and that increase reduces interest-sensitive spending. At the lower bound, this particular offset is absent while the constraint binds. Fiscal expansion can therefore provide more support to output than under a policy rule that raises rates, all else equal.
Worked Example: Comparing a Lower-Bound Response With a Rate Increase
A hypothetical economy has \(c=0.60\) and \(b=0.40\). A government-purchases increase creates a fiscal impulse \(g=0.80\) percentage points of potential output. Compare the output-gap change at a binding lower bound with the change when the central bank follows \(\Delta i=0.50x\), where \(\Delta i\) is in percentage points and expected inflation is fixed.
Method and assumptions. Use the reduced-form demand equation. At the lower bound, the nominal rate does not change. In the comparison case, the rate responds to output according to the stated rule. There are no other demand or supply changes.
At the lower bound. Since \(\Delta r=0\), the equation is \(x=0.60x+0.80\). Subtracting \(0.60x\) from both sides gives \(0.40x=0.80\), so \(x=2.00\) percentage points.
The multiplier is \(1/(1-0.60)=2.50\). The first-round fiscal impulse is \(0.80\), and the induced demand is \(0.60(2.00)=1.20\). Together, \(0.80+1.20=2.00\), as calculated.
With the rate response. Fixed expected inflation means \(\Delta r=\Delta i=0.50x\). Substitute into the model:
Thus \(0.60x=0.80\), giving \(x=1.3333\ldots\) percentage points. The policy rate rises by \(0.50(1.3333\ldots)=0.6667\ldots\) percentage points. The rate-related demand reduction is \(0.40(0.6667\ldots)=0.2667\ldots\), and induced demand is \(0.60(1.3333\ldots)=0.80\). The net is \(0.80+0.80-0.2667\ldots=1.3333\ldots\). Rounded to two decimal places, the output-gap changes are 2.00 and 1.33 percentage points.
Common mistake avoided: applying the lower-bound multiplier in both cases. When the rate rises, the term \(b\phi x\) must be included; omitting it would overstate the output response.
Monetary Accommodation and Real Interest Rates
Monetary accommodation means that monetary policy does not offset the added demand by raising interest rates. When the lower bound binds, the central bank has no room to cut the nominal rate further, but the rate can still remain at its bound rather than rise. In the basic model, that absence of a rate increase prevents the interest-rate crowding-out channel from reducing the fiscal effect.
There can also be a real-rate channel. Since the real interest rate is approximately \(r=i-\pi^e\), where \(i\) is the nominal rate and \(\pi^e\) is expected inflation, expected inflation can rise while the nominal rate remains constrained. The real rate then falls. A lower real rate may support interest-sensitive consumption and investment. This additional effect depends on expectations actually changing; it is not guaranteed merely because the nominal rate is at its lower bound.
Worked Example: An Increase in Expected Inflation at the Lower Bound
In a hypothetical economy, the nominal policy rate is fixed at its lower bound of \(0.25\%\). Expected inflation initially is \(1.25\%\). Assume \(c=0.50\), \(b=0.40\), and a fiscal impulse \(g=0.84\) percentage points of potential output. For this example only, assume expected inflation rises by \(0.20\) percentage points for each one-percentage-point increase in the output gap: \(\Delta\pi^e=0.20x\). Calculate the output response, accounting for the real-rate channel.
Method and assumptions. Keep the nominal rate fixed and use the stated relationship between the output gap and expected inflation. First find the change in the real rate, then substitute it into the demand equation. The proportional expectation relationship is a hypothetical modeling assumption, not a general law.
The initial real rate is \(0.25\%-1.25\%=-1.00\%\). Because the nominal rate stays fixed, \(\Delta r=-\Delta\pi^e=-0.20x\) percentage points. Substituting into the demand model gives:
Therefore \(0.42x=0.84\), and \(x=2.00\) percentage points. Expected inflation rises by \(0.20(2.00)=0.40\) percentage points, from \(1.25\%\) to \(1.65\%\). The real rate falls from \(-1.00\%\) to \(0.25\%-1.65\%=-1.40\%\), a decrease of \(0.40\) percentage points.
The three contributions in the equation are the fiscal impulse \(0.84\), induced demand \(0.50(2.00)=1.00\), and the real-rate contribution \(-0.40(-0.40)=0.16\). Their sum is \(0.84+1.00+0.16=2.00\). If expected inflation were fixed instead, the response would be \(0.84/(1-0.50)=1.68\) percentage points. Rounded to two decimal places, the response with the expectation channel is 2.00 percentage points.
Common mistake avoided: treating a \(0.25\%\) nominal rate as a \(0.25\%\) real rate. The real rate must account for expected inflation; here it is negative even before the fiscal expansion.
How Much the Lower-Bound Advantage Depends on Crowding Out
The difference between the two multipliers is largest when a rate increase would otherwise have a substantial effect on demand. In the model, that requires both a responsive monetary policy (\(\phi>0\)) and interest-sensitive spending (\(b>0\)). If the central bank would not raise rates, or if borrowing costs have little effect on spending, the rate-offset channel is small.
The model also abstracts from factors that may limit fiscal transmission. Some spending may go to imports rather than domestic output; households may save part of transfers; firms may face supply constraints; and policy implementation can take time. Earlier material on fiscal-policy lags and crowding out addresses those issues in more detail. The simple multiplier comparison isolates the effect of the interest-rate constraint, not the total effect of every real-world policy.
Worked Example: Separating Fiscal Feedback From Rate Crowding Out
A hypothetical economy has \(c=0.40\), \(b=0.50\), and a fiscal impulse \(g=1.50\) percentage points of potential output. At a binding lower bound, the policy rate does not rise. For comparison, outside the bound, the central bank follows \(\Delta i=0.80x\), and expected inflation is fixed. Calculate both output responses and identify the rate-related crowding out in the second case.
Method and assumptions. Solve the same reduced-form demand equation in each policy regime. The direct fiscal impulse and other model parameters are identical, so the difference comes from the rate response.
At the lower bound. With \(\Delta r=0\), \(x=0.40x+1.50\). Thus \(0.60x=1.50\), giving \(x=2.50\) percentage points. Induced demand is \(0.40(2.50)=1.00\), and \(1.50+1.00=2.50\).
With a rate increase. Fixed expected inflation gives \(\Delta r=0.80x\). Then:
So \(x=1.50\) percentage points. The policy rate rises by \(0.80(1.50)=1.20\) percentage points. The rate-related demand reduction is \(0.50(1.20)=0.60\), while induced demand is \(0.40(1.50)=0.60\). The net is \(1.50+0.60-0.60=1.50\). Rounded to two decimal places, the output responses are 2.50 and 1.50 percentage points.
Common mistake avoided: reporting the rate increase itself as the amount of crowding out. The rate rises by 1.20 percentage points, but its demand effect is \(b\Delta r=0.60\) percentage points in this model.
Interpretation and Exam Traps
- Do not equate the lower bound with a zero real rate. Expected inflation determines the difference between the nominal and real rates.
- Do not assume a larger multiplier is automatic. The comparison depends on the policy response, interest sensitivity, and other model assumptions.
- Keep rate units consistent. If \(b\) is defined per percentage-point change in the real rate, use percentage points in \(b\Delta r\), not a decimal rate such as \(0.01\) for one percentage point.
- Separate direct and induced effects. The fiscal impulse is \(g\); income-induced demand is \(cx\); interest-rate crowding out is \(b\Delta r\).
The central result is that fiscal expansion can be especially useful for demand stabilization when the policy rate cannot be reduced further: the usual rate-increase offset may be absent, and a fall in the real rate may add support if expected inflation rises. These channels explain a potential increase in effectiveness, not a guarantee of a particular output or inflation outcome.
Check Your Understanding
Use the short-run demand model and the stated assumptions in each problem. Give output-gap changes in percentage points.
- At a binding lower bound, \(c=0.70\) and a government-purchases increase creates \(g=0.60\). Expected inflation is fixed. Calculate the output response. (Answer: 2.00 percentage points)
- Let \(c=0.50\), \(b=0.25\), and \(g=1.00\). The real interest rate rises by 1.00 percentage point independently of output, and expected inflation is fixed. Calculate \(x\) from \(x=cx+g-b\Delta r\). (Answer: 1.50 percentage points)
- A central bank follows \(\Delta i=0.50x\) outside the lower bound. Let \(c=0.60\), \(b=0.40\), and \(g=0.60\), with expected inflation fixed. Find the output response and the policy-rate increase. (Answer: output response 1.00 percentage point; rate increase 0.50 percentage point)
- At the lower bound, the nominal rate is fixed. Expected inflation rises from \(1.00\%\) to \(1.50\%\), and \(b=0.30\). What is the real-rate change and the added demand contribution from this change? (Answer: real rate falls by 0.50 percentage point; demand contribution is 0.15 percentage point)
- Compare a binding lower bound with a rate-response rule. Let \(c=0.40\), \(b=0.50\), \(g=0.90\), and, outside the lower bound, \(\Delta i=0.40x\). Expected inflation is fixed. Calculate both output responses and the rate increase outside the bound. (Answer: lower-bound response 1.50 percentage points; response outside the bound 1.125 percentage points; rate increase 0.45 percentage point)