By popular demand
Today I thought I would start with giving kudos to the ECB in general and ECB-researchers in particular: You do write a whole lot of solid and really interesting analysis! Based on what I hear from our diligent support-team, a recent remake of a sliver of Athansios Orphanides work (with extra everything) – the “Orphanides rule” – has caught the eye of quite many students of economics.
A swift introduction: The Orphanides rule is a simple monetary policy rule in the spirit of the original Taylor rule (which we also have had a brief look into). The main difference is that it is analyzed on a first-difference and it also does away with many (not all) unobservable variables (output gaps, NAIRU etc) often employed in traditional Taylor-rules. Instead, the Orphanides-rule uses deviations of the one-year ahead forecasts of inflation (𝐸𝜋𝑡+1) from target (𝜋̅) and of the one-year ahead GDP-growth forecast (𝐸∆𝑦𝑡+1) from an estimate of potential GDP-growth (∆𝑦̅):
∆𝑖 = 0.5(𝐸𝜋𝑡+1 − 𝜋̅) + 0.5(𝐸∆𝑦𝑡+1 − ∆𝑦̅)
The problem, of course, is how to choose which expectations and potential/target should go into the calculations. I figure that ECB’s ditto would be ideal, but as the referred to paper showed, the combo of one-year ahead forecasts from the ECB’s Survey of Professional Forecasters (SPF) and the European Commissions estimate for potential GDP-growth seems to give quite a good “fit” as well (we’re talking eye-ball econometrics of a higher degree here). I have used a few alternatives in the graphs below, and if it were easier to find longer-term data on the ECB’s and ESCB’s staff macroeconomic projections (MEP’s) I would indeed have included those too (we’re looking into it!). That said, if you open the chart in the MB- application you can, of course, substitute for any data you like (even the currently quite short data on MEP-projections), perhaps even add some of your own.
Looking at an OLS-estimated version (incl in the MB-file), you can see that the fit is quite alright with an R2 above 0.7 and other statistics also reasonably well-behaved. Nonetheless, given the number of possible combinations of data on expectations, targets, potential etc., I think looking at a range of estimates could be informative on how to judge the monetary policy stance.
Do note the wide range for most of the post-2010 period, when the effective lower bound was also binding (hence, I have also added the Wu & Xia shadow rate to the graph). What really sticks out, nonetheless, are developments over the last couple of years, where it is suggested that actual monetary policy (Wu & Xia) has been more expansionary than what any of the monetary policy rules deemed necessary. And, not only that, due to recent downward revisions of inflation and growth expectations, all policy rules indicate that the ECB’s “window of opportunity to hike” (if there even is such a thing) has all but passed. Negative interest rates are here to stay.