For over three decades, inflation targeting has served as the dominant intellectual framework for monetary policy design in advanced economies. Introduced by New Zealand in 1990 and adopted subsequently by Canada, the United Kingdom, Sweden, and eventually embraced in spirit by the Federal Reserve, the regime rests on a deceptively simple premise: announce a numerical inflation target, commit credibly to achieving it, and let the transparency of that commitment do much of the stabilization work through its effect on expectations.

The theoretical case, formalized most rigorously in Michael Woodford's Interest and Prices, demonstrated that an explicit nominal anchor could resolve the indeterminacy problems inherent in earlier monetary frameworks. When the central bank commits to a well-defined rule—or at minimum a systematic reaction function—private agents can form expectations that are self-reinforcing, reducing the output cost of disinflation and improving the variance trade-off between inflation and real activity.

Yet the framework that delivered the Great Moderation now confronts challenges its architects did not fully anticipate. The effective lower bound on nominal interest rates, persistent supply-side shocks, and mounting evidence on distributional heterogeneity have prompted serious reconsideration. Average inflation targeting, nominal GDP targeting, and price-level targeting have all entered the policy conversation. Understanding where inflation targeting succeeded—and where its assumptions strain—is essential for designing whatever comes next.

The intellectual case for inflation targeting derives from the rational expectations revolution and the time-inconsistency problem first articulated by Kydland and Prescott. Without an institutional commitment mechanism, policymakers face a persistent temptation to exploit the short-run Phillips curve, generating inflation surprises that temporarily boost output. Over time, the private sector adjusts its expectations upward, and the economy settles at higher inflation with no permanent output gain—the classic inflationary bias.

An explicit numerical target addresses this problem through two channels. First, it provides a coordination device for expectations formation. When firms, households, and financial markets share a common understanding of the central bank's objective, the dispersion of inflation expectations narrows. In New Keynesian models with Calvo-style price setting, this reduced dispersion directly lowers the sacrifice ratio associated with any given monetary policy adjustment. The target becomes a focal point in the game-theoretic sense.

Second, the target creates an accountability framework that substitutes for the kind of rule-based commitment that is difficult to implement literally. Central banks that miss their target must explain the deviation publicly—through inflation reports, testimony, or open letters to government. This institutional architecture imposes reputational costs that partially replicate the commitment solution to time inconsistency without requiring the rigid adherence to a mechanical rule that would be suboptimal in the face of unforeseen shocks.

Woodford's contribution was to show that under certain conditions, a credible inflation target renders the rational expectations equilibrium determinate. In models where monetary policy is specified only as an interest rate rule without a nominal anchor, multiple equilibria—including sunspot-driven fluctuations—can arise. The target pins down the price level path, eliminating self-fulfilling expectational dynamics. This is not merely an academic nicety; it provides the theoretical justification for why communication about objectives matters as much as the instrument setting itself.

Crucially, the anchoring effect depends on credibility, which is endogenous. A newly announced target without institutional backing or track record does little to shift expectations. The theoretical models predict—and the empirical evidence largely confirms—that the benefits of inflation targeting accumulate over time as the central bank demonstrates its willingness to act consistently with the announced objective, even when doing so is politically costly.

Takeaway

An explicit inflation target does not merely constrain the central bank—it restructures the entire expectations equilibrium, making the private sector an active participant in delivering the policy outcome rather than an obstacle to it.

The empirical literature on inflation targeting's track record is extensive, and the headline finding is broadly favorable. Cross-country studies by Ball and Sheridan, Gonçalves and Salles, and subsequently by the IMF staff have consistently found that countries adopting explicit inflation targets experienced lower and more stable inflation compared to their pre-targeting periods. The average level of inflation fell, the variance of inflation declined, and importantly, the persistence of inflation shocks—a key measure of expectational anchoring—diminished substantially.

The more contentious question is whether these gains reflect the targeting framework itself or broader global trends. Ball and Sheridan's influential 2005 study argued that much of the improvement could be attributed to regression to the mean: countries that adopted inflation targeting tended to be those with higher initial inflation, and their convergence to lower rates mirrored trends in non-targeting countries. This critique has force, but subsequent work exploiting variation in adoption timing, institutional design, and exposure to common shocks has generally found a residual treatment effect, particularly in emerging market economies where the credibility deficit was largest.

The performance during the Great Recession provides a particularly informative test. Inflation targeting central banks generally maintained well-anchored long-run inflation expectations even as headline inflation became volatile due to commodity price movements and the financial crisis. Survey-based and market-based measures of expected inflation five to ten years ahead remained remarkably stable in targeting regimes—a finding consistent with the theoretical prediction that a credible target should insulate long-run expectations from transitory shocks.

On output stability, the evidence is more nuanced. The reduction in output volatility during the Great Moderation occurred in both targeting and non-targeting economies, making attribution difficult. However, flexible inflation targeting regimes—those explicitly accounting for output stabilization alongside inflation—appear to have achieved more favorable variance trade-offs than strict targeting implementations. The distinction matters: it suggests that the framework's value lies not in single-minded inflation focus but in the disciplined flexibility it enables.

Where the record is less convincing is in the framework's handling of financial stability. Neither inflation targeting's theory nor its institutional implementation anticipated the buildup of systemic financial risk that characterized the pre-2008 period. Price stability, it turned out, was necessary but not sufficient for macroeconomic stability. This gap has driven the integration of macroprudential considerations into central bank mandates—an evolution that sits uneasily with the clean theoretical elegance of the original inflation targeting paradigm.

Takeaway

Inflation targeting demonstrably improved inflation outcomes and expectational anchoring, but its track record reveals a framework optimized for demand-driven fluctuations that struggles when the dominant shocks are financial or supply-side in nature.

The tension between strict and flexible inflation targeting has been present since the framework's inception, but it has intensified as policymakers confront a more complex macroeconomic environment. Strict inflation targeting—targeting inflation alone at all horizons—is a theoretical benchmark that no central bank has ever literally implemented. Even the most inflation-focused institutions, such as the Reserve Bank of New Zealand in its early years, acknowledged that responding to output and employment fluctuations was operationally necessary to avoid excessive instrument volatility.

Lars Svensson formalized the distinction by characterizing flexible inflation targeting as the minimization of a loss function that places weight on both inflation deviations from target and output gap variability. The optimal policy under this framework involves a deliberate trade-off: following a supply shock that raises inflation and depresses output simultaneously, the central bank returns inflation to target gradually, accepting temporary deviations to avoid amplifying real economic costs. The Federal Reserve's dual mandate can be interpreted as a particularly flexible variant of this approach, with employment stabilization receiving explicit and roughly equal weight.

The Federal Reserve's 2020 shift to average inflation targeting represented the most significant conceptual evolution within the framework. By committing to offset past undershoots with future overshoots, the Fed attempted to address the asymmetric constraint imposed by the effective lower bound. The logic is straightforward in theory: if agents expect the central bank to allow temporarily above-target inflation after a recession, real interest rates decline even when the nominal rate is at zero, providing additional stimulus precisely when it is most needed. In practice, the framework was stress-tested almost immediately by the post-pandemic inflation surge, and its credibility implications remain actively debated.

More radical alternatives have gained intellectual traction. Nominal GDP level targeting would automatically prescribe easier policy when real output falls and tighter policy when the economy overheats, without requiring the central bank to separately estimate the output gap—a quantity that is measured with notorious imprecision. Price-level targeting shares the make-up property of average inflation targeting but in a more transparent form. Each alternative addresses specific limitations of standard inflation targeting, but each also introduces new communication challenges and transition risks.

The deepest challenge may be distributional. Heterogeneous agent models—HANK frameworks in particular—demonstrate that monetary policy operates through channels that affect different households very differently. Interest rate changes redistribute between borrowers and savers, between asset holders and wage earners. A framework that targets aggregate inflation may be optimal in a representative agent sense yet generate welfare losses for substantial population segments. Whether inflation targeting can be adapted to accommodate these insights, or whether an entirely different conceptual architecture is needed, is the frontier question in monetary policy design.

Takeaway

The evolution from strict to flexible to average inflation targeting reveals a framework progressively absorbing the complexity it initially assumed away—raising the question of how much flexibility an 'inflation target' can accommodate before it ceases to function as an anchor.

Inflation targeting earned its dominance through a rare combination of theoretical coherence and practical effectiveness. It solved the credibility problem that plagued discretionary monetary policy, provided a transparent accountability structure, and delivered measurably improved macroeconomic outcomes across diverse institutional settings.

But the framework was designed for a world of demand-driven business cycles, well-behaved Phillips curves, and representative agents. The post-2008 landscape—characterized by the effective lower bound, persistent supply disturbances, and growing attention to heterogeneity—has exposed assumptions that once seemed innocuous. Each adaptation stretches the original architecture further from its clean theoretical foundations.

The question facing monetary economics is not whether inflation targeting failed—it largely succeeded on its own terms. The question is whether its terms remain adequate for the policy challenges ahead, or whether the next generation of frameworks must be built on fundamentally different conceptual ground.