McCallum Rule: Formula and Uses
Table of Contents
The McCallum rule asks how quickly a central bank’s monetary base would need to grow to keep nominal income near a specified path. It offers a different way to organize monetary-policy analysis from a rule that sets an interest rate. The useful comparison separates three questions: what outcome policymakers want, which instrument they adjust, and how that instrument reaches spending and prices.
This article explains the level-gap rule in Bennett McCallum’s 1987 proposal, calculates a hypothetical benchmark, and compares it with the Taylor rule and nominal GDP targeting. The historical calibration is an educational example, not a recommendation for today’s Federal Reserve. U.S. operating-framework descriptions were checked in November 2026.
What Is the McCallum Rule?
The McCallum rule is a proposed monetary-base instrument rule: it specifies base growth using a target for nominal income, a correction for changes in base velocity, and feedback when nominal income misses its target. The monetary base consists of currency in circulation plus reserve balances. It differs from the public’s bank deposits and broader money supply measures. [1] [2]
The rule adjusts the monetary base to pursue a nominal-income objective. A base-growth number is an instrument setting, not a direct forecast of real growth, inflation or bank lending.
McCallum’s original 1987 article used nominal gross national product, or GNP. Modern discussions often use nominal GDP instead. Both are nominal-income measures, but their economic coverage differs; substituting one changes an empirical exercise. The worked example below explicitly uses hypothetical GDP while retaining the 1987 rule’s functional form. [1] [3]
The rule also differs from constant money growth. Its instrument responds to conditions according to a prespecified formula. A policy can therefore be responsive while remaining rule based. The broader monetary policy framework explains the institutions and tools behind that distinction.
Start With a Target Nominal-Income Path
Nominal income combines quantities produced with their prices. A target path specifies a desired level for each date, rather than merely asking whether this quarter’s growth is strong. To construct it, choose both an initial target level and a growth rate.
For the 1987 proposal, McCallum used approximately 3% annual nominal-GNP growth, matching his assumed 3% long-run real growth and aiming at approximately zero long-run inflation. That was a historical design choice. A different trend-growth estimate or inflation objective changes the nominal-income path; 3% is not a universal policy constant. [1]
The quarterly natural-log equivalent of 3% ordinary annual growth is g* = ln(1.03) / 4 = 0.0073897, rounded to 0.00739 in the original equation. This is approximately 0.739% quarterly log growth. Entering 0.03 as a quarterly decimal-log target would change the rule substantially.
A nominal-income objective does not specify how each quarter’s income growth divides between real production and inflation. That division depends on economic conditions and adjustment processes. The quantity theory of money supplies useful background, but an accounting identity alone does not establish how a policy change will affect either outcome.
The McCallum Rule Formula and Timing
The following is an equivalent, compact presentation of equation (2) in McCallum’s 1987 paper, using his proposed 0.25 response coefficient. Let B denote the monetary base, X nominal income, and V = X / B its base velocity. Lowercase letters are their natural logs: b = ln(B), x = ln(X), and v = x − b. A star denotes the chosen target. [1]
Δbt = bt − bt−1
Since v = x − b, the numerator of the velocity term can also be written xt−1 − xt−17 − bt−1 + bt−17. The two endpoints span 16 intervals, or four years. This is not a one-quarter velocity change or a four-quarter average.
| Term | Meaning | Convention |
|---|---|---|
| Δbt | Change in the log base from quarter t − 1 to t | Quarterly decimal log growth |
| g* | Growth of the chosen nominal-income path | ln(1.03)/4 for the historical 3% calibration |
| v̄t | Average historical base-velocity growth | 16 quarterly intervals through t − 1 |
| xt−1* − xt−1 | Log of target income divided by actual income | A level gap at the same date |
| 0.25 | Response to the log level gap | Coefficient in this version, not an annual growth percentage |
For GDP data reported at annual rates, use the same convention for actual and target levels and throughout the velocity history. A consistent annualization factor cancels from log differences and target/actual ratios. Mixing conventions does not. Historical replications also need the specified base series and adjustments, rather than automatically substituting a modern unadjusted total-base series. [3]
Why the Signs Point in These Directions
- Higher target nominal growth raises base growth, with the other terms fixed.
- Faster velocity growth lowers base growth. If nominal income grows more rapidly relative to the base, the rule subtracts that historical drift from the required base expansion.
- Income below its path raises base growth. Target divided by actual income exceeds one, making its log positive. Income above the path makes this feedback negative.
The feedback changes growth relative to the other terms. An above-target economy does not necessarily imply a shrinking base. Nor does the 0.25 coefficient mean that the income gap will close by exactly one-quarter next quarter; that outcome requires a model of policy transmission.
For a log shortfall of 0.01, feedback adds 0.25 × 0.01 = 0.0025 to quarterly log base growth: 0.25 quarterly log percentage point, or 1 annualized log percentage point. This is approximately the response to an ordinary 1% shortfall; ordinary percentage gaps and log gaps are close but not identical.
A Worked McCallum Rule Example
Consider a hypothetical next-quarter benchmark, using nominal GDP in place of the original nominal GNP measure. Assume the previous quarter’s target GDP was $10.0 trillion, actual GDP was $9.8 trillion, and the monetary base was $2.0 trillion. All GDP levels use the same annual-rate convention. Assume base velocity rose 4% in total over the preceding 16 quarterly intervals. Keep the historical 3% annual target path and 0.25 level-gap coefficient.
Actual GDP is 2% below target: (10.0 − 9.8) / 10.0 = 0.02. The rule uses the log ratio instead. The following decomposition keeps all contributions in quarterly decimal-log units.
| Component | Calculation | Contribution |
|---|---|---|
| Target path growth | ln(1.03) / 4 | +0.00738970 |
| Velocity adjustment | −ln(1.04) / 16 | −0.00245129 |
| Level-gap feedback | 0.25 × ln(10.0 / 9.8) | +0.00505068 |
| Δbt | Sum, using unrounded values | 0.00998908 |
The positive shortfall feedback more than offsets the velocity subtraction. With the same velocity trend and GDP exactly on target, feedback would be zero and quarterly log base growth would instead be 0.00738970 − 0.00245129 = 0.00493841. This comparison isolates the effect of the income miss.
Convert the Log Change to a Base Level
A log change is not quite an ordinary percentage change. Convert it by exponentiating:
Quarterly ordinary growth = exp(0.00998908) − 1 = approximately 1.0039%.
Bt = $2.0 trillion × exp(0.00998908) = approximately $2.02008 trillion.
The target base is approximately $20.08 billion higher than the previous quarter’s $2.0 trillion level. Calculations retain unrounded log inputs before displaying these results.
That $20.08 billion is a change in the hypothetical base target, not a calculated securities-purchase requirement. Currency demand, other central-bank liabilities and offsetting balance-sheet movements can affect how a base target relates to operations.
Annualize Without Mixing Definitions
Multiplying the quarterly decimal log change by 400 gives 3.9956% annualized log growth. The exact effective annual growth rate, if this same quarterly rate were repeated four times, is 100 × [exp(4 × 0.00998908) − 1] = approximately 4.0765%. These are different reporting conventions. Neither says that the rule will choose the same setting in each subsequent quarter.
McCallum Rule vs Taylor Rule
The Taylor rule sets a short-term nominal interest rate, while the McCallum rule above sets base growth. McCallum’s primary-author comparison describes Taylor’s original illustrative form as follows, with rates in annual percentages and the real-output gap in percent. [3]
The first 2 is an assumed 2% equilibrium real rate; the second 2 is a 2% inflation target. Those are historical illustrative assumptions, not current estimates. Taylor’s original illustration used four-quarter GDP-deflator inflation and real GDP relative to an estimated trend. Inflation appears both on its own and inside the inflation gap. Omitting the standalone inflation term would change the nominal-rate rule.
For a separate hypothetical Taylor calculation, assume inflation is 3% and the output gap is −2%. The result is 2 + 3 + 0.5 × (3 − 2) + 0.5 × (−2) = 4.5%. This is an annual nominal-rate benchmark under those assumptions. It cannot be directly compared with the McCallum example’s base-growth percentage as if they were the same instrument.
| Question | 1987 McCallum version | Original Taylor calibration |
|---|---|---|
| Instrument | Quarterly monetary-base growth | Nominal short-term interest rate |
| Feedback | Lagged nominal-income level relative to a path | Inflation relative to target and real output relative to its reference |
| Distinct input challenge | Base-velocity drift, base definition and initial income path | Equilibrium real rate and estimated real-output gap |
| Common challenge | Data revisions, timing and uncertain transmission | Data revisions, timing and uncertain transmission |
Neither rule removes judgment. The McCallum version avoids inserting a current real-output gap directly, but its target path still needs a view about long-run real growth and the inflation objective. A meaningful comparison evaluates both rules in a consistent economic model and operating environment, rather than declaring one superior because its equation looks simpler.
McCallum Rule vs Nominal GDP Targeting
Nominal GDP targeting identifies an outcome to pursue. It does not, by itself, specify whether the central bank adjusts an interest rate, the monetary base or another instrument. McCallum’s base rule supplies a particular operating formula. His later comparison also examines an interest-rate rule responding to nominal-income growth, showing that the instrument and objective can be paired differently. [3]
Also distinguish a level target from a growth target. A level target retains a path and responds to accumulated misses. A growth target can restore normal growth while leaving an earlier shortfall in place. The ECB’s framework discussion makes this distinction explicitly. [4]
In a separate hypothetical annual example, suppose the target index is 100 at Year 0 but actual nominal GDP is 98. Let both subsequently grow 3% per year, without assuming any response to the McCallum rule.
| Date | Target index | Actual index |
|---|---|---|
| Year 0 | 100.00 | 98.00 |
| Year 1 | 103.00 | 100.94 |
| Year 2 | 106.09 | 103.97 |
For example, Year 1 actual income is 98 × 1.03 = 100.94, while target income is 100 × 1.03 = 103. Actual remains 2% below target. A rule responding to the level gap still detects a shortfall even though current growth matches the target growth rate.
This distinction matters when comparing published McCallum equations. His 2000 growth-gap version uses a 0.5 response to the difference between target growth and the previous quarter’s nominal-income growth. It differs from the 1987 level-gap version and its 0.25 coefficient explained here. Mixing those coefficients, gaps or timing conventions creates a different rule. [3]
Why Ample Reserves and IORB Matter
The 1987 formula does not describe the Federal Reserve’s current operating procedure. The New York Fed’s implementation guidance describes an ample-reserves framework, in which administered rates, especially interest on reserve balances or IORB, steer the federal funds rate. IORB is the rate the Fed pays on eligible reserve balances. [5] [6]
When reserves remain in the ample region, modest changes in their quantity need not materially change the federal funds rate. The return available on reserves influences banks’ willingness to lend in money markets; other standing facilities support rate control. IORB alone is not a firm floor for every participant, because some cannot earn it. [7]
This changes how to interpret a base-growth benchmark. A larger reserve quantity with the administered-rate settings held fixed is not automatically the same monetary easing that an older quantity-based model might imply. The New York Fed distinguishes reserve-management purchases that support rate control from purchases intended to ease financial conditions or repair market functioning. [5]
Bank reserves are distinct from customer deposits. A commercial-bank loan creates a deposit; additional reserves do not force a fixed multiple of new loans. Credit demand, capital, liquidity, profitability and risk matter. The Bank of England’s explanation of modern money creation makes this distinction clear. [8]
The implication is to evaluate the rule’s transmission assumptions, rather than treating reserve quantity as irrelevant. A substantial reduction that makes reserves scarce can create funding pressure. Asset purchases can also affect longer-term financial conditions through other channels. Our monetary policy transmission mechanisms guide explains those channels; yield curve control illustrates a different operating commitment tied directly to selected bond yields.
How to Evaluate a McCallum Benchmark
- Name the version. Record the source equation, level or growth feedback, coefficient, velocity window and lag convention.
- Specify the income path. State the initial level, growth objective, GDP or GNP measure, and why that target is appropriate for the question.
- Match the data. Keep frequency, seasonal treatment, annual-rate conventions and base definitions consistent. A historical adjusted-base series and a current total-base series need reconciliation.
- Respect the information set. A t − 1 label does not ensure that the final revised observation was available when a decision was made. Use release vintages for a real-time historical test.
- Separate a benchmark from a simulation. Applying a rule to observed data produces a conditional setting. Simulating adoption requires modeling how the economy and future rule inputs respond.
- Check the operating framework. Examine reserve remuneration, liquidity demand and the link from the instrument to spending before interpreting a larger base target as a particular stance.
The rule is useful for making assumptions visible and comparing policy frameworks. It does not by itself forecast a bond yield, inflation release or currency return. For the broader inputs behind currency analysis, see forex fundamental analysis.
Common Mistakes
- Using an annual number in a quarterly equation. Convert ordinary annual growth to quarterly logs before combining terms.
- Confusing a level gap with a growth gap. Normal growth can coexist with income below its original path.
- Reversing the velocity sign. Rising historical velocity is subtracted; falling velocity contributes positively through that subtraction.
- Using M2 velocity with a base rule. The denominator must be the monetary aggregate specified by the rule.
- Treating 0.25 as a current policy constant. It belongs to the selected historical version and units.
- Reading base expansion as guaranteed lending or inflation. The equation supplies a setting; economic outcomes depend on behavior and the operating regime.
Limitations of the McCallum Rule
The velocity correction is deliberately slow. A four-year average smooths short-run fluctuations but can lag a structural change in the demand for base money. Interest paid on reserves, payment practices and regulation can change the relationship being estimated. An identity linking nominal income, the base and velocity does not guarantee that a backward-looking correction stabilizes future income.
The target path is a substantive choice. Its starting point and long-run growth assumption affect measured misses. GDP revisions can change the apparent gap after a decision. Nominal-level targeting also depends on credible communication about responding to past shortfalls; that credibility cannot be created by writing an equation. Cleveland Fed remarks identify these measurement and commitment problems. [9]
Model evidence is conditional. McCallum’s original paper investigated performance in estimated models. That supports evaluating a proposal, not asserting that the same coefficients will work under every financial structure or crisis. A simple rule also leaves many practical questions about financial stress and implementation outside its formula. Treat its output as a benchmark whose assumptions can be examined, with any proposed policy use requiring much fuller analysis. [1] [3]
Frequently Asked Questions
Sources
- Bennett T. McCallum: The Case for Rules in the Conduct of Monetary Policy: A Concrete Example, September/October 1987. Printed pp. 12–14, equation (2), target path, quarterly units, velocity window and proposed coefficient.
- Federal Reserve Board, via FRED: Monetary Base: Total. Series notes defining the monetary base.
- Bennett T. McCallum: Alternative Monetary Policy Rules, Winter 2000. Section 1, Taylor calibration, different growth-gap McCallum version, units and timing; Section 2, limits of conditional historical comparisons.
- Vítor Constâncio, ECB: The future of monetary policy frameworks, May 25, 2017. Nominal GDP level versus growth targeting.
- New York Fed: Monetary Policy Implementation. Current operating framework, administered rates and distinct asset-purchase purposes.
- Federal Reserve Board: Interest on Reserve Balances FAQs. IORB and reserve-balance definitions.
- Ihrig, Senyuz and Weinbach, Federal Reserve: Implementing Monetary Policy in an Ample-Reserves Regime: The Basics, July 1, 2020. Reserve demand and administered-rate mechanism.
- McLeay, Radia and Thomas, Bank of England: Money creation in the modern economy, March 14, 2014. Banking mechanics and limitations of the mechanical money-multiplier description; U.K. institutional context.
- Loretta J. Mester, Cleveland Fed: Monetary Policy Frameworks, January 5, 2018. Nominal-level targeting, starting points, revisions and credibility.
Sources accessed November 15, 2026 UTC. All worked numbers are independently calculated hypothetical scenarios, not current U.S. policy prescriptions or replicated historical simulations.
Disclaimer
This article is for educational and informational purposes only and does not constitute investment or policy advice. Historical policy-rule calibrations and hypothetical calculations do not predict inflation, interest rates, currency movements or investment returns. Actual policy effects depend on economic conditions, institutions, data and behavior. Consult a qualified financial professional about your circumstances.
