Taylor (1993) Discretion versus Policy Rules in Practice

Summary

This is the paper that introduced the Taylor rule. Taylor argues that econometric policy-evaluation research (rational-expectations macro, the Lucas critique, time-inconsistency, credibility) has produced a consensus that responsive policy rules dominate discretion, and that good rules call for adjusting the short-term interest rate in response to inflation and real output. His central device is a specific, "hypothetical but representative" rule — $r = p + 0.5y + 0.5(p-2) + 2$ — that closely approximates actual Federal Reserve policy over 1987–1992. The larger message is that the concept of a policy rule should be preserved as a benchmark and discipline even though no central bank can (or should) mechanically follow an algebraic formula: judgment is still needed for special episodes and for interpreting the data. He illustrates with two case studies — German unification and the 1990 oil-price shock. (Carnegie-Rochester Conference Series on Public Policy 39: 195–214.)

Key Claims

• Rules beat discretion — the research consensus. Modern macro (post–Lucas critique) finds that traditional discretionary policy evaluation was flawed, that rational expectations does not imply policy ineffectiveness, that credibility has real benefits, and that time-inconsistency makes rules superior to discretion. The preferred rules are responsive, not fixed (not constant money growth): they move the instrument in response to the price level and real income.
• The Taylor rule. The representative rule is $$r = p + 0.5\,y + 0.5\,(p - 2) + 2,$$ where $r$ is the federal funds rate, $p$ is inflation over the previous four quarters, and $y = 100\,(Y-Y^*)/Y^*$ is the percentage deviation of real GDP from trend/target. Equivalently $r = 1.5\,p + 0.5\,y + 1$: the funds rate responds to inflation with coefficient 1.5 and to the output gap with coefficient 0.5.
• Implied targets and the equilibrium rate. The rule assumes an inflation target of 2% and an equilibrium real rate of 2%: when both inflation and output are on target ($p=2$, $y=0$), the funds rate equals 4% (2% real). The funds rate rises whenever inflation exceeds 2% or output exceeds trend.
• It fits recent Fed policy. The rule tracks actual Federal Reserve behavior closely over roughly 1987–1992, suggesting the Fed was, in effect, following a rule of this form even without announcing one.
• Rules can't be followed mechanically — so preserve the concept. Purely algebraic rules cannot encompass everything: judging whether a price rise is temporary or permanent needs multiple price measures and expectations data; measuring potential output needs forecasts of productivity, participation, and the natural rate; and special episodes (the October 1987 crash, when the Fed injected reserves) require discretion. But that these arguments "sound like" the case for discretion does not justify abandoning rules — the rule should be kept as the baseline from which deviations are justified.
• Case studies. German unification and the 1990 oil-price shock are used to show how a rule-based framework operates when confronted with real, non-routine shocks.

Concepts Introduced or Extended

• Taylor Rule
• Monetary Policy Shocks
• Monetary Neutrality

Entities Mentioned

• John B. Taylor

Quotes

"One policy rule that captures the spirit of the recent research and which is quite straightforward is: $r = p + .5y + .5(p - 2) + 2$ … where $r$ is the federal funds rate, $p$ is the rate of inflation over the previous four quarters, [and] $y$ is the percent deviation of real GDP from a target."

"It is important to preserve the concept of a policy rule even in an environment where it is practically impossible to follow mechanically the algebraic formulas economists write down to describe their preferred policy rules."

My Take

Few three-parameter formulas have been as influential: the Taylor rule became simultaneously a descriptive benchmark (does actual policy look like this?), a normative prescription (should it?), and — crucially for this wiki — an identifying restriction for the monetary-policy equation in structural VARs and DSGE models. What is easy to forget from citations is Taylor's own framing: the paper is at least as much about the limits of mechanical rules as about the rule itself — he insists the rule is a disciplining baseline, not an algorithm, and that judgment about temporary-vs-permanent shocks and potential output is unavoidable. The "1.5 on inflation" is also the empirical seed of the later Taylor principle ($\phi_\pi>1$ for determinacy) that underlies the good-policy reading of the Great Inflation debated by Sims–Zha, Primiceri, and Clarida–Galí–Gertler.