Alternative Monetary Policy Paths and Downside Risk

IMF News, October 29, 2018

Source details

Canonical URL

Alternative Monetary Policy Paths and Downside Risk

Other formats

• Markdown version
• Structured JSON version
• Bundle manifest

Bibliographic details

• Authors: Tobias Adrian
• Published: October 29, 2018

Core argument

• Economic forecasts that provide point estimates for the conditional mean of GDP growth (or the GDP gap) ignore risks around the central forecast and can be overly optimistic.
• Policy makers' focus on risks around the forecast has increased; central banks commonly discuss the balance of risks to growth and inflation.
• Three advances justify incorporating endogenous forecast distributions into practical monetary policy:
  • Density forecasting is now relatively straightforward to implement (Adrian, Boyarchenko, Giannone 2018); financial conditions powerfully forecast the whole future GDP distribution and financial vulnerability manifests as downside risk to output.
  • A parsimonious New Keynesian model with financial vulnerability (Adrian and Duarte 2018) matches these forecast densities, generates conditional distributions, and delivers closed-form optimal policy rules as a function of financial vulnerability.
  • The monetary policy process should consider the endogenous evolution of risk in alternative policy paths alongside impacts on inflation and the output gap.

Conditional density forecasting of output and inflation

• Two-step estimator (Adrian, Boyarchenko, Giannone 2018):
  • First: regress future output (growth or gap) on current output and current financial conditions using quantile regressions to estimate conditional quantiles rather than conditional means.
  • Second: fit a flexible distribution function across quantiles at every point in time to generate a conditional time series of output densities.
• Estimator performance:
  • Excellent out-of-sample performance relative to naïve random walk estimators, more complicated Bayesian approaches, or fully parametric maximum likelihood approaches.
  • Robust and straightforward to implement in practice.

Stylized facts from Adrian, Boyarchenko, Giannone (2018) and Growth-at-Risk

• Stylized facts:

1) The conditional volatility of future GDP is predicted primarily by financial conditions. 2) The one quarter and one year ahead conditional density of GDP evolves primarily through changes in the conditional mean and conditional volatility; conditional skewness and kurtosis vary only slightly. 3) The conditional mean and conditional volatility are negatively correlated, producing an unconditional GDP distribution strongly skewed to the left.

• Growth-at-Risk (measured by the lower 5th percentile of future GDP growth conditional on financial conditions):
  • The term structure of Growth-at-Risk is consistent with endogenous risk-taking and the volatility paradox.
  • Periods of easy financial conditions: short-run low downside risks and high expected growth, but a medium-term reversal with large increases in downside risks.
  • Times of tight financial conditions: large near-term downside risks that diminish in the medium term.
• Empirical scope:
  • Current estimate of the term structure of Growth-at-Risk shown across 11 advanced economies.
  • Prevailing easier-than-average financial conditions help limit near-term downside risks relative to mid-range financial conditions over the past 40 years, but suggest larger downside risks two years from now and as soon as in one year if the economy is in a credit boom.

A New Keynesian model with financial vulnerability (NKV)

• Model formulation and motivation:
  • Suggests modeling endogenous second moments (volatilities of the output gap as functions of state variables) as a parsimonious source of nonlinearity.
  • Reduced form closely related to standard first-order New Keynesian (NK) models but with error terms featuring endogenous volatilities depending on state variables.
  • The NKV model is a micro-founded New Keynesian model with financial vulnerability solved in a fully nonlinear fashion; its reduced form is a standard NK model with endogenous second moments that are linear functions of state variables.
• Key properties and implications:
  • The model does not suffer from certainty equivalence: risk matters and conditional second moments are not constant (data strongly reject constant second moments).
  • The NKV model allows analytical solution to the optimal policy rule.
  • Policy prescriptions relative to the standard Taylor rule:
  • During booms (output gap closed or positive and inflation at or above target), the optimal rule is slightly tighter than the standard Taylor rule—leaning slightly against buildup of financial vulnerabilities during easy financial conditions.
  • During times of a large output gap, monetary policy is looser than under a standard Taylor rule to induce risk taking and help close the output gap more quickly.
• Cross-country calibration and welfare:
  • Model calibrated to 21 economies (advanced and emerging) shows that welfare gains from using the optimal policy rule that conditions on financial vulnerability can be sizable.
  • These welfare gains are achieved while assuming a central bank with an inflation-targeting mandate; the resulting optimal rule is flexible inflation-targeting that conditions on financial vulnerability.

Alternative policy paths with endogenous risk

• Reduced-form NKV discrete-time model (Adrian, Duarte, Liang, Zabczyk 2018):
  • Extends the standard NK three-equation model by adding financial vulnerability as an additional state variable (four equations total).
  • Introduces two new parameters:
  • One parameter determines the slope of conditional output volatility (measured by financial conditions) as a function of expected output gap—this is the “vulnerability channel” and pins down how vulnerability depends on expected growth.
  • The second parameter governs how expected growth depends on the pricing of risk and appears in the IS curve—this is the “financial accelerator” channel of BGG.
  • Both parameters are micro-founded in Adrian and Duarte (2018).
• Comparing NKV-optimal rule versus standard 1993 Taylor rule:
  • Example calibration: start with inflation roughly at target, output gap closed, and financial conditions very loose.
  • NKV rule prescribes initial policy tightening to tighten financial conditions and increase output volatility more sharply in the near term than the Taylor rule. Once financial conditions become less loose, monetary policy becomes less restrictive and output and inflation gaps close.
  • The crossing of financial condition paths illustrates an intertemporal risk tradeoff: higher near-term conditional volatility may yield lower medium-term volatility.
  • Symmetry: when output and inflation gaps are closed and financial conditions are tight, NKV rule prescribes looser policy than the Taylor rule to increase risk taking in the short term.
• Decision framework and caveats:
  • The “right decision” depends on policymakers’ preferences: discount factor, risk aversion, intertemporal elasticity of substitution.
  • Real-time measurement imprecision of output and inflation gaps, and desire for more realistic medium-sized DSGE models, means judgment is required in practice.
  • Communication of objectives, measurement, and time horizons also matters.

Policy implications and macroprudential interaction

• Monetary policy that actively mitigates vulnerability, in addition to focusing on inflation and the output gap, will reduce negative skewness of the output gap over time; prudent monetary policy in the presence of endogenous risk mitigates downside risks to GDP.
• Macroprudential policy:
  • Overall safety and soundness should be addressed via structural prudential tools (micro- and macroprudential) which determine resilience to shocks.
  • Cyclical macroprudential tools can mitigate downside risks to GDP; the NKV framework can model monetary and cyclical macroprudential policies jointly.
  • Macroprudential tools and governance have limitations; pronounced cyclical variation in downside risks persists.
  • Macroprudential tools usually take several quarters, and in some cases a year or more, to phase in; monetary policy decisions are implemented immediately.
• Current practice:
  • Tracking risks to the forecast aligns with current monetary policy conduct: committees receive regular financial stability briefings and consider the balance of risks.
  • References to risks by monetary policy makers have been increasing and some appear to react to shifts in financial conditions.

Conclusion — three-step case for endogenous risk in policy models

• 1) Empirical finding: conditional volatility of output depends on financial conditions and creates an intertemporal risk-return tradeoff for policymakers; this finding holds across countries and time.
• 2) Modeling: endogenous risk can be parsimoniously modeled in an NKV model with risk as an additional endogenous state variable and only two additional parameters relative to the standard NK model (“financial vulnerability” and “financial accelerator”); the model is analytically tractable and captures the stylized facts.
• 3) Policy analysis: this framework provides endogenous forecast distributions; the NKV model generates highly skewed and fat-tailed unconditional output distributions even with Gaussian shocks. Policymakers considering alternative policy paths should also consider the future path of risk.

IMF Communications Department, “Alternative Monetary Policy Paths and Downside Risk,” speech as prepared for delivery, October 29, 2018.

References

• Tobias Adrian
• United States and the IMF
• Speeches
• PRESS CENTER
• The Term Structure of Growth at Risk
• https://www.imf.org/en/home