The Science of Monetary Policy: A New Keynesian Perspective

Questions & answers

Q1. What is the paper’s stated purpose, and why append “New Keynesian Perspective” to the title?

The paper “summarizes what we have learned from this recent research on monetary policy,” organizing the discussion around a simple theoretical model built up from a stripped-down baseline case to more realistic complications, in order to “assess how the predictions from theory square with policy-making in practice” (Section 1, p. 2-3). The authors are explicit that their approach “is based on the idea that temporary nominal price rigidities provide the key friction that give rise to non-neutral effects of monetary policy,” and append “New Keynesian Perspective” to signal that they combine Keynesian nominal rigidities with “the recent methodological advances in macroeconomic modeling” from dynamic general-equilibrium theory, distinguishing their approach from real-business-cycle theory (which rejects nominal rigidities) and from frameworks emphasizing frictions in money demand instead (Section 1, p. 3).

Q2. What is the baseline model, and why build it “from first principles”?

A simple two-equation New Keynesian model – a forward-looking IS-type output-gap equation and a forward-looking Phillips curve linking inflation to expected future inflation and the output gap – derived from optimizing household and firm behavior rather than assumed as reduced-form relationships, “because we are interested in characterizing policy rules in terms of primitive factors” (Section 2, p. 3-4). The authors stress an important departure from the classical Tinbergen-Theil targets-and-instruments problem: because both the output gap and inflation depend on expectations of future policy (not just current policy), “the target variables depend not only on the current policy but also on expectations about future policy,” which is what makes the credibility of monetary policy relevant to the analysis (Section 2.3, p. 17-18).

Q3. What is “Result 3,” and how does the paper use it to evaluate historical U.S. policy?

“Under the optimal policy, in response to a rise in expected inflation, nominal rates should rise sufficiently to increase real rates… the coefficient on expected inflation should exceed unity” (Section 3, p. 26) – the paper’s formal statement of what is now commonly called the Taylor principle. The authors apply it directly as “a very simple criteria for evaluating monetary policy,” citing their own companion work (Clarida, Gertler and Gali 1997) to argue that “U.S. monetary policy in the pre-Volcker era of 1960-1979 violated this strategy… Federal Reserve policy tended to accommodate rather than fight increases in expected inflation. Nominal rates adjusted, but not sufficiently to raise real rates,” while “[s]ince 1979… the Federal Reserve appears to have adopted the kind of implicit inflation targeting strategy that equation (3.6) suggests,” systematically raising real rates in response to anticipated increases in inflationary expectations (Section 3, p. 26).

Q4. What is “Result 4,” and what is its economic logic?

“The optimal policy calls for adjusting the interest rate to perfectly offset demand shocks… but perfectly accommodate shocks to potential output… by keeping the nominal rate constant” (Section 3, p. 26-27). The logic follows from the structure of the baseline model: a demand shock moves output and inflation without altering the natural level of output, so it can be fully neutralized at no cost to the inflation-output trade-off; a shock to potential output, by contrast, does not by itself require any change in the policy rate under the baseline specification, since it does not open up an inflation-output tension the way a cost-push/supply shock would.

Q5. What is the paper’s central new result on commitment, and how does it differ from the traditional inflationary-bias story?

Result 7 states that “if price-setting depends on expectations of future economic conditions, then a central bank that can credibly commit to a rule faces an improved short run trade-off between inflation and output. This gain from commitment arises even if the central bank does not prefer to have output above potential” (Section 4, p. 39). This is explicitly distinguished from the traditional Kydland-Prescott/Barro-Gordon argument, in which the gain from commitment comes from eliminating an inflationary bias that arises only if the central bank’s output target exceeds the natural rate: “in contrast to the traditional analysis, this gain from commitment is not tied to the desire of the central bank to push output above potential, but to the forward-looking nature of inflation… in our baseline model” (Section 4, p. 38-39). The mechanism is that a central bank without commitment “would like to convince the private sector that it will be tough in the future” while avoiding contracting demand today, but a rational private sector anticipates the temptation to renege, so only a credible commitment secures the more favorable expected path and, correspondingly, a larger nominal-rate response to expected inflation than under discretion (Section 4, p. 38-40).

Q6. What is the “opportunistic” approach to disinflation, and when is it optimal?

Building on a proposal by Alan Blinder from his time at the Federal Reserve Board, the opportunistic approach holds that when inflation is above but near the optimum, “policy should not contract demand. Rather, it should take an ‘opportunistic’ approach… waiting until achieving the inflation target could be done at the least cost in terms of incremental output reduction” – i.e., waiting for favorable supply shocks to bring inflation down rather than actively tightening (Section 5.3, p. 58-59). Formalizing this via a non-smooth objective function that penalizes small output deviations more than small inflation deviations, the authors derive Result 12: “If there is more cost associated with small departures of output from target than with small departures of inflation, then an opportunistic approach to disinflation may be optimal. This policy, further, is equivalent to targeting inflation around a zone as opposed to a particular value” (Section 5.3, p. 61) – what Bernanke and Mishkin (1997) term “inflation zone targeting.” The authors note this distinct opportunistic behavior only arises when cost-push (supply-side) shocks are present in inflation, since demand shocks alone make opportunistic and conventional gradualist rules equivalent (Section 5.3, p. 60-61).

Q7. What complications does the paper introduce beyond the frictionless baseline model, and what do they imply?

Section 5 works through imperfect information and lags, large money-demand shocks, model/parameter uncertainty, and non-smooth preferences, deriving further numbered results (e.g., Result 9 on imperfect information, Result 10 on money-demand volatility, Result 11 on parameter uncertainty attenuating the policy response) (Section 5, p. 45-58). Among the practical implications the authors draw out: these complications help “makes clear why modern central banks (especially the Federal Reserve Board) have greatly downgraded the role of monetary aggregates in the implementation of policy” (Section 1, p. 5-6), and they motivate why central banks smooth interest-rate adjustments even though the baseline optimal-policy results do not call for such smoothing on their own (Section 5, “Interest Rate Smoothing,” discussed p. 55-58) – an issue the concluding section flags as still unresolved (Section 8, p. 79).

Q8. What open questions does the paper flag for future research?

Four in particular: better understanding the determinants and persistence of inflation (arguing, citing Galí-Gertler 1998 and Sbordone 1998, that real marginal cost/unit labor costs rather than the output gap may be the right forcing variable in the Phillips curve); extending the closed-economy analysis to open-economy settings (exchange-rate regime choice, policy coordination, CPI versus domestic inflation targeting); understanding policy at the zero lower bound on nominal rates, already a live concern given conditions “in Japan” at the time of writing; and explaining why central banks smooth interest-rate changes in practice (Section 8, “Concluding Remarks,” p. 78-79). On the zero-bound question specifically, the authors note that “when the nominal rate is at zero, the only way a central bank can reduce the real interest rate is to generate a rise in expected inflation,” calling how a central bank should engineer this, and whether fiscal cooperation is required, “important open questions” (Section 8, p. 79).