Rules vs Discretion in Monetary Policy: Time Inconsistency, Inflation Temptation, and the Role of Reputation

Under discretion, the monetary authority has room to generate inflation higher than what the public initially expects. Ex post, surprise inflation can appear “beneficial” for instance by stimulating real activity through an expectations-augmented Phillips curve or by reducing the real burden of nominal liabilities. Yet once the public understands policymakers’ incentives and forms expectations rationally, systematic surprise inflation cannot persist in equilibrium. The outcome is a higher average inflation (and money growth) than the “ideal” benchmark, while the surprise component itself disappears. This is the classic logic behind the superiority of rules (commitment) over discretion—and why reputation can act as an imperfect substitute for formal commitment.

Executive Summary

Core problem: time inconsistency—policies that are optimal when announced are no longer optimal when implemented because policymakers have incentives to “cheat” expectations.
Under discretion: rational expectations internalize incentives → equilibrium inflation is higher, with no systematic surprise inflation.
Under an ideal rule: inflation can be reduced (in the simplest case, close to zero), but the ideal rule may fail to be enforceable when it relies on reputation alone.
Reputation yields an intermediate equilibrium: outcomes lie between discretion and the ideal rule; the more “impatient” the policymaker (higher discounting), the closer the outcome is to discretion.
Policy implication: credibility is an asset; institutional designs that strengthen commitment and credible punishment reduce average inflation without relying on surprise.

Academic Notes

These notes summarize key intuitions and results; full derivations and step-by-step modeling are provided in the Methods section.

1) Key concept: surprise inflation, rational expectations, and time inconsistency

Barro–Gordon’s key insight: policymakers may gain ex post from inflation exceeding expected inflation ( $\pi_t>\pi_t^e$ πt>πte) because output temporarily rises. But with rational expectations, the public anticipates this incentive, so in discretionary equilibrium average inflation rises while equilibrium surprise inflation is zero ( $\pi_t=\pi_t^e$ ).

Brief model illustration: $y_t-\bar y=\alpha(\pi_t-\pi_t^e),\ \alpha>0$

A political/target pressure can be represented by $y^*>\bar y$ (gap $k>0$ ).

2) Policy objective: “inflation costs” vs “surprise benefits”

The paper uses a minimal cost–benefit structure:

Inflation costs rise (quadratically) with $\pi_t$ .
Surprise benefits are tied to $\pi_t-\pi_t^e$ .

Minimal per-period representation: $\ell_t=\frac{a}{2}\pi_t^2-b_t(\pi_t-\pi_t^e),\quad a>0,\ b_t>0$

Higher $b_t$ bt means stronger incentives to engineer surprises.

3) Three regimes: discretion, the ideal rule, and cheating

(i) Discretion: the policymaker sets inflation after expectations are formed. With rational expectations, equilibrium features positive inflation bias but no systematic surprise.

(ii) Ideal rule (commitment): if commitment is enforceable, inflation can be pushed down (simplest benchmark: near zero), eliminating the inflation bias.

(iii) Cheating: if the public believes the rule (low $\pi^e$ πe), the policymaker is tempted to inflate above the rule to exploit surprise benefits. This looks attractive ex post but cannot persist once expectations adjust.

4) Reputation as enforcement: temptation vs enforcement

In a repeated interaction, reputation disciplines deviations: cheating today raises expected inflation tomorrow, making future outcomes worse. A rule is sustainable only if: $\text{Temptation (one-shot gain)} \le \text{Enforcement (discounted future loss)}$ Temptation (one-shot gain)≤Enforcement (discounted future loss)

With discount factor $\beta$ , a shorter horizon (lower $\beta$ ) weakens reputational enforcement.

5) Main result: the best enforceable rule and “intermediate” inflation

When the ideal rule is not enforceable, the equilibrium converges to a best enforceable rule, producing inflation outcomes between discretion and the ideal rule.

Comparative statics:

Higher $\beta$ (more patient policymaker) → more rule-like outcomes → lower inflation.
Lower $\beta$ (more impatient/short horizon) → more discretion-like outcomes → higher inflation.

6) Emergency states and shifting incentives

When the marginal benefit of surprises $b_t$ increases recessions, wars, fiscal stress, large nominal debt temptation rises, rules become harder to sustain, and inflation pressures increase.

7) “Bite the bullet”: investing credibility

In state-contingent versions, policymakers may choose tighter actions in “good times” to invest credibility, preserving room to maneuver in emergencies without losing the expectations anchor.

Relevance: political pressure & electoral horizons

The Barro–Gordon reputational framework offers a formal language for monetary policy dynamics under rising political pressure especially around elections or fiscal crises. When policymakers’ horizons shorten (effectively lowering $\beta$ ), reputational considerations weaken, pushing outcomes toward discretion: higher average inflation and less credible commitment. Conversely, institutions that lengthen horizons clear mandates, operational independence, transparency, and consistent communication strengthen reputation as an enforcement device, making policy more rule-like and inflation expectations better anchored.

References

Barro, R. J., & Gordon, D. B. (1983). Rules, Discretion and Reputation in a Model of Monetary Policy (NBER Working Paper No. 1079). RULES, DISCRETION AND REPUTATION
Kydland, F. E., & Prescott, E. C. (1977). Rules Rather than Discretion: The Inconsistency of Optimal Plans. Journal of Political Economy.