Central bank credibility

For neoclassical economists, central bank credibility means avoiding high inflation rates degenerating into low economic growth and high unemployment rates (Barro and Gordon, 1983). A credible central bank fulfils its low inflation announcement, and agents believe its commitment to price stability. A central bank's credibility is therefore measured by the difference between the central bank's inflation plan and what the public believes about these plans (Cukierman, 1992), or, in the framework of inflation targeting, by the gap between inflation expectations (or current inflation) and the inflation target (Svensson, 2011).

Theoretical foundations of the neoclassical view of central bank credibility (see Barro and Gordon, 1983) are the vertical Phillips curve, the rational expectations hypothesis and the game-theoretic approach of time inconsistency: the central bank has private information on its type ("hawk" or "dove" on inflation) and plays a game against economic agents. A non-credible central bank plays a fooling game by violating its announced inflation target. It fools agents' expectations to exploit the Phillips curve and boost employment, the cost being higher inflation ("inflation bias").

For neoclassical economists, the central bank makes these non-credible announce- ments, because it is controlled by politicians trying to be re-elected by reducing employ- ment temporarily. In this sense, a central bank lacks credibility when it is not independent from politicians. For neoclassical economists, this credibility problem gave rise to the "Great Inflation" in the 1970s, despite its unlikelihood as being the only cause of that phenomenon.

For neoclassical economists, credibility matters because a credible central bank benefits from a credibility "effect" or "bonus": the output cost of a central bank's disinflationary policy is reduced, and more generally the economic situation and monetary policy efficiency improve. For neoclassical economists, the central bank can obtain this credibility bonus by announcing a credible disinflationary policy stance, and only the announce- ment of a quick disinflation ("cold-turkey" disinflation) is credible (Sargent, 2013). The "credibility effect" of this announcement is that agents believe in this announced disinflation policy and quickly adjust their expectations accordingly. This adjustment of agents' expectations confers a credibility "bonus": a credible central bank benefits from a disinflationary "free lunch" (at no cost to the economic system).

According to some members of the new neoclassical synthesis, the benefits of cred- ibility go further (see Svensson, 2011). A credible central bank anchors inflation expec- tations, which consequently do not react to shocks. Therefore, owing to the credibility "bonus", the impact of shocks on the economy is attenuated, and the central bank has flexibility to respond to shocks with no fear of inflating expectations. It follows the ulti- mate bonus of credibility: the inflation/output trade-off is dampened, and so are long- term interest rates, because the inflation-related risk premium vanishes.

Nevertheless, empirical evidence does not fully support these credibility "effects", which are not necessarily offered by agents' expectations (Hutchison and Walsh, 1998). In the real world, disinflation is usually not a "free lunch": the cost of quick disinflation is measured by a "sacrifice ratio"; that is, higher unemployment rates and/or output losses generated as a result of disinflation policy. Also, in practice an inflation-focused credibility does not provide enough flexibility for a central bank to deal with large shocks and related inflation/output trade-offs. Inflation-focused credibility can even constrain central bank flexibility to respond to shocks, causing unnecessary social losses. A credibility/flexibility trade-off emerges thereby (Walsh, 2010).

In the mid 1990s the neoclassical credibility view of central banking was criticized, because time inconsistency was judged as outdated. A central bank's credibility is thus no longer related to the political temptation to exploit the Phillips curve, but is often confused with transparency: a credible central bank has a clear communication and is therefore predictable. Blinder (2000) clarifies the notion of credibility, defining it as "matching deeds to words": the public believes the announcement of the central bank. The correct terminology in this regard, however, is "confidence", not "credibility" (Carré and Le Héron, 2006).

In the early 2000s, New Keynesians revived credibility, even if they did not retain the inflation-bias hypothesis. For them, contrary to neoclassical economists, credibility is not a game against but with agents' expectations. Credibility is not a fooling game but a communication-signalling game. Communication therefore becomes a monetary policy instrument fostering (inflation, output, interest-rate) expectations by publishing a central bank's forecasts of these variables.

For neoclassical economists, this credibility was unreliable, because communication was "cheap talk", macroeconomic variables were persistent, and agents' expectations were partly backward-looking. By contrast, New Keynesians defended (and continue to defend) forward-looking models with a credible central bank managing inflation expectations and controlled the current rate of inflation depending on its expected rate. Credibility becomes thereby "expectations management" (Woodford, 2003).

Under inflation targeting, central bank credibility is defined as the gap between infla- tion expectations and the official inflation target. In this regard, the underlying hypoth- esis by New Keynesians is still rational expectations, but with the New-Keynesian Phillips curve. For them, as for neoclassical economists, credibility is crucial not per se, but because of its positive effects on the whole economic system, notably the reduced costs of maintaining low inflation rates and reduced economic fluctuations. Yet these benefits are model-dependent, and diminish with non-forward-looking agents.

New Keynesians envisage imperfect credibility when agents doubt the central bank's commitment to its inflation target, giving rise to imperfectly managed expectations. Besides, the Bank for International Settlements (see Borio and White, 2003) shows a "credibility paradox": by stabilizing the rate of inflation too much, the central bank gives agents a false sense of security, leading them to more risky behaviour that could degener- ate into financial instability.

Monetarists, neoclassical economists, and more recently New Keynesians (Woodford, 2003) propose a pre-commitment to a monetary policy rule, because they judge non- credible a central bank under discretion (no commitment). In the real world, a rule valid during a large shock does not exist: de facto the commitment is conditional. For Rogoff (1985), credibility requires a personal commitment with a "conservative" central banker more inflation-averse than the whole society. This, however, generates a democratic deficit with a loss function different from society's. Credibility can also come from an institu- tional commitment to low inflation rates, via total central bank independence or putting the inflation target in the Constitution. The cost of this is an unaccountable central bank, and reduced flexibility with a quasi-unamendable Constitution. Credibility can be earned via a reputation or a long history of low inflation rates. The cost is a central bank trading- off low economic growth against low inflation rates to build its anti-inflation reputation. For New Keynesians, incentive-compatible mechanisms of delegation, a contract about the inflation target, or a dismissal rule on the central banker à la Walsh (1995) can also build credibility.

The Great Recession challenged the anti-inflation-oriented credibility of central banks. During financial crises, the monetary authority can be credible by announcing that it will do everything possible, even generating inflation to avoid deflation (Blinder, 2012). In fact, inflation is only one aspect of central banking: the US Federal Reserve, for instance, was created in 1913 to fight financial instability. Moreover, the financial crisis that burst in 2008 empirically confirms that a central bank's pre-commitment to a rule is not credible.

See also:

Central bank independence; Credibility and reputation; Financial crisis; Financial insta- bility; Friedman rule; Inflation; Inflation targeting; Phillips curve; Rules versus discre- tion; Taylor rule; Time inconsistency.