Discretion versus policy rules in practice

Questions & answers

Q1. How does the paper redefine “policy rule” to make the concept usable in practice?

Taylor explicitly broadens the definition of a policy rule beyond either a fixed setting for policy instruments or a mechanical algebraic formula, to include a “systematic” policy implemented with judgment by policymakers who understand the rule’s underlying logic but recognize that operating it “requires judgment and cannot be done by computer.” He proposes distinguishing three separate policy-rule issues — the design of a rule, the transition to a new rule once designed, and the day-to-day operation of a rule once in place — arguing that apparently discretionary policy actions can often be more usefully interpreted as transitions between rules or as judgment-based operation of an existing rule, rather than as a rejection of rules altogether.

Q2. What does multicountry model research say about the design of good monetary policy rules?

Comparisons across nine multicountry rational-expectations models (the Bryant-Hooper-Mann volume) find that interest-rate rules responding to deviations of the price level and real output from target outperform rules that target the money supply or the exchange rate, though there is no consensus on the exact size of the response coefficients. Taylor’s own multicountry model comparisons of fixed versus flexible exchange-rate regimes for the G-7 find that a flexible exchange-rate system produces both lower output variability (the standard deviation of output nearly doubles under fixed rates in Germany and Japan) and lower price variability (Japan and Germany show more than twice the price volatility under a fixed exchange rate pegged to the dollar) — leading him to conclude the flexible-rate system dominates fixed rates on both dimensions for the countries considered, and that placing some weight on real output (not just the price level) in the interest-rate rule generally improves performance over a pure price rule.

Q3. What is the specific policy rule Taylor proposes, and how well does it match actual Fed behavior?

The proposed rule is r = p + 0.5y + 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 percent deviation of real GDP from its trend (2.2% per year over 1984:Q1-1992:Q3) — implying the funds rate should rise when inflation exceeds a 2% target or when output exceeds trend, and should equal 4% (a 2% real rate plus 2% inflation) when both are exactly on target.** Despite using deliberately round-number coefficients chosen “for easy discussion” rather than statistically optimized values, this hypothetical rule tracks the actual federal funds rate path remarkably closely from 1987 to 1992, with the most significant deviation occurring in 1987 when the Fed eased policy in response to the stock market crash rather than following the rule’s implied path.

Q4. Why does the paper argue that transitions between policy rules require special treatment, distinct from either steady-state rule operation or pure discretion?

Standard rational-expectations analysis of a policy rule assumes agents have already adjusted their expectations and behavior to a long-standing, credible rule, but Taylor argues this assumption is unrealistic immediately after a new rule (e.g., a lower inflation target) is adopted — agents instead likely form expectations partly by studying policymakers’ track records and assessing the new policy’s credibility, so the short-run impact of a new rule can differ substantially from what a standard rational-expectations analysis would project. He adds a second, independent reason for gradual transitions: real economic rigidities (long-term wage contracts, investment commitments, loan contracts) prevent instantaneous behavioral adjustment, so transitions — such as a disinflation — should generally be phased in gradually and announced publicly to let people unwind prior commitments without large losses.

Q5. What two practical approaches does the paper propose for incorporating rule-like behavior into actual policymaking without mechanical formula-following?

The first approach is to explicitly add a specific algebraic rule (such as the paper’s example) as one additional input among the many the FOMC already considers, e.g., having Fed staff include in briefing materials how recent decisions compare with the rule’s prescription or forecasts of the funds-rate path the rule implies under several coefficient variants. The second approach, which Taylor likens to patent law’s establishment of general principles while leaving implementation details to patent officials and courts, is to codify only the qualitative properties of a good rule (the 1990 Economic Report of the President’s description that the Fed “generally increases interest rates when inflationary pressures appear to be rising and lowers interest rates when…recession appears to be more of a threat” is offered as an example) — leaving the magnitude of any response to expert judgment about factors such as the interest sensitivity of aggregate demand.

Q6. How does the 1990 oil-price shock case study illustrate a principled departure from the rule’s literal prescription?

Following Iraq’s invasion of Kuwait on August 2, 1990, oil prices roughly doubled and Council of Economic Advisers/model estimates suggested a temporary 50% oil-price increase could raise the GDP deflator by about 1% and (with a lag) reduce real output by a comparable amount; taken literally, the policy rule would call for raising interest rates in response to the price-level increase, but Taylor argues this would have been inappropriate because the shock was recognized as temporary. The key evidence was that oil futures prices barely moved (the December 1991 futures price rose only about $4/barrel while the spot price rose $25), and subsequent analysis confirmed increased production elsewhere could largely offset the lost Iraqi/Kuwaiti supply if the conflict ended — so central banks in most countries, including the U.S., held interest rates on the path they would have followed absent the shock, a deliberate deviation from the rule’s literal reading that Taylor treats as consistent with, not contrary to, sound rule-based policymaking; on the fiscal side, the Gramm-Rudman-Hollings law initially would have forced offsetting spending cuts that blocked the automatic stabilizers from cushioning the shock, and the budget law’s deficit targets were subsequently revised to let the stabilizers operate.

Q7. How does the German unification case study illustrate distinguishing real from inflationary sources of an interest-rate movement?

U.S. and German long-term interest rates rose sharply in early 1990, which under the policy rule’s logic could signal rising inflation expectations calling for a rate increase — but Taylor argues the evidence instead pointed to a real, unification-driven shift in German investment demand: the West German government budget swung from a 0.2%-of-GDP surplus in 1989 to a 3-4%-of-GDP deficit in 1990-91, consistent with anticipated unification-related capital demand, and forward-looking model simulations found that plausible increases in German capital demand could account for roughly a one-percentage-point rise in real interest rates. Because the rate increase was attributable to real investment-demand factors rather than a change in inflation expectations, Taylor concludes interest-rate policy did not need to be tightened in response — again framing a departure from the naive literal signal (rising long rates) as the economically correct, rule-consistent response once the underlying cause is correctly diagnosed.

Q8. What is the paper’s overall conclusion about the practical role of policy rules?

Taylor concludes that policy rules should be understood as broad frameworks for “systematic and credible” policymaking rather than literal formulas to be mechanically applied, and that clarifying the distinct concepts of rule design, transition, and operation — along with either explicitly referencing specific rules or codifying their qualitative properties in policy deliberations — offers a practical path for incorporating rule-like discipline into actual central-bank decision-making. The two case studies are presented as concrete illustrations that principled judgment in applying a policy rule (correctly diagnosing whether a shock is temporary or real) is compatible with, and indeed required by, genuinely rule-based policymaking, rather than being evidence that rules should be abandoned in favor of unconstrained discretion.