At their meeting in Sydney last month the G20 announced a promising new “G20 Growth Agenda.” They centered the initiative on a goal of raising real GDP by over $2 trillion (see communique paragraph 3) compared with current projections by increasing economic growth by ½ percent per year for 5 years, with a commitment to come back next time with proposed policies to achieve the goal.

Where do these numbers come from?  The IMF staff simulated an econometric model of the G20 countries and showed that growth would speed up by this amount if the countries adopted certain policies.  Most of the policies considered by the IMF are supply-side (structural) rather than demand side policies.  They include “fiscal consolidation,”  “product market reforms,”  “labor participation reforms,”   “other labor market reforms,” along with some “infrastructure” and “rebalancing” reforms that have a longer run impact through saving and investment.   According to the IMF “about 1/3” of increase in growth rates comes from “positive productivity spillovers between members, the main source of growth spillovers.”

Of course, the heavy lifting will be at future meetings when the countries are supposed to come with their own list of pro-growth policies that they are pursuing, presumably similar to those suggested in the IMF simulations.  Will this actually happen?  How can it be encouraged?

We can learn from a very similar initiative of ten years ago: the 2003 G7 Agenda for Growth. It too focused on pro-growth supply-side policies, with countries committing to particular policies. The commitment process lasted for a couple of years but eventually lost steam—as is so often the case with such international initiatives—as people rotated out of various leadership positions in their governments. The new G20 Growth Agenda is likely to suffer the same fate unless there a way to make the initiative longer lasting than the current leadership positions. Let’s hope the G20 and the international financial institutions can create the needed commitment mechanism.