Wednesday, July 18, 2012

The Optimal Rate Of Inflation

We’re on a optimisation binge today. After the last post on international reserves, here’s a piece on the optimal level of inflation (abstract; emphasis added):

How Inflation Affects Macroeconomic Performance: An Agent-Based Computational Investigation
Quamrul Ashraf, Boris Gershman, Peter Howitt

We use an agent-based computational approach to show how inflation can worsen macroeconomic performance by disrupting the mechanism of exchange in a decentralized market economy. We find that increasing the trend rate of inflation above 3 percent has a substantial deleterious effect, but lowering it below 3 percent has no significant macroeconomic consequences. Our finding remains qualitatively robust to changes in parameter values and to modifications to our model that partly address the Lucas critique. Finally, we contribute a novel explanation for why cross-country regressions may fail to detect a significant negative effect of trend inflation on output even when such an effect exists in reality.

So 3 percent it is. Malaysia’s record is not quite there:

01_cpi

For much of the 1990s and pre-1985 for instance, inflation was too high. For the past decade, complications like the commodity price boom have otherwise detracted from a decent inflation record. Nevertheless, since the beginning of the Great Moderation in 1981, Malaysia’s average inflation rate is right on the money – 3 percent.

One thing to note about this paper is that its one of the very, very few economics papers I’ve seen to use agent-based modelling, which is more commonly used in physics but has been touted as a new way forward for econometrics – you can read more about the methodology here.

Technical Notes

Ashraf, Quamrul and Boris Gershman & Peter Howitt, "How Inflation Affects Macroeconomic Performance: An Agent-Based Computational Investigation", NBER Working Paper No. 18225, July 2012

1 comment: