In thinking about the future of monetary policy, it’s important to consider legislative reforms and appointments, but it’s also important to consider the economic models that have come to be a key part of policy making in central banks. The Bank of Canada showed a great deal of vision last week when it invited economists and practitioners to discuss “Central Bank Models: The Next Generation.”

We can learn from history here. In my keynote address for the conference I reviewed the 80-year history of macro policy models from Jan Tinbergen’s model of 1936 to the present. I showed how policy making with models began in “path-space” (simulating paths for the policy instruments and seeing the impact on the paths for target variables), but, with a major paradigm shift 40 years ago, moved to “rule-space” (simulating rules for the instruments and seeing the impact of economic stability over time).  Central bank models followed these developments, though with a lag, to the benefit of monetary policy and economic performance.

However, there was a retrogression in parts of the central banking world in the past dozen years, and economic performance has deteriorated with a great recession and a very slow recovery. This history suggests that a pressing problem for central bank research is to get back to a “rule-space” framework. The framework was good while it lasted, and still benefits countries that continue to implement it. The departure from this approach should be fundamentally reconsidered, which means finding out how to get back to the framework, understanding why the departure occurred, and figuring out how to prevent future departures.

But what can researchers do? What kind of research can help?

I offered some some ideas for research direction and workflow, giving examples from Chapter 15 by Volker Wieland, Elena Afanasyeva, Meguy Kuete, Jinhyuk Yoo and Chapter 28 by Jesper Linde, Frank Smets, Raf Wouters  in the new Handbook of Macroeconomics.  As explained here, research should focus on

— how changes in the economy and models of the economy affect monetary policy rules.  Examples include changes in the technology of financial inter-mediation, increased integration of the financial and real sides of the economy, behavioral economics factors, new distribution channels of monetary policy, agent-based models, heterogeneous price and wage setting, and the impact of an effective lower bound on the interest rate.

— designing models for the purpose of evaluating policy rules. Having a purpose in mind helps determine the size, scope and type of model.

— robustness through transparent and replicable macro model comparisons such as in the new Macroeconomic Model Comparison Initiative using the Macro Model Data Base.

— the interface between policy models and policy decisions. Here more transparent reporting on policy rules used in practice as in recent legislation that has passed the House would be useful.

— the connection between monetary policy rules and an international rules-based monetary system. Perhaps there is no more important, and no more difficult, application of “rule-space” analysis.  One research idea would study an international agreement in which each central bank would report and commit to its monetary rule or strategy, and thereby help build the foundation of the international monetary system.

— distinguishing between instrument rules, “forecast targeting,” and “constrained discretion,” which means delving deeper into the classic rules versus discretion debate.

In my view, these ideas are worth pursuing at central banks and by monetary researchers outside of central banks, even if one has different views of the reasons for poor economic performance during the past decade.