The OECD’s recommendations for the European Central Bank

 

 

In its survey on the euro area of January 5, the OECD devotes significant space to a discussion of monetary policy, especially the role of money in monetary policy. See also our updated briefing note this subject. The OECD argues that recent empirical evidence weakened two of the rationales for the use of monetary indicators by central banks: money demand functions were less stable recently than during the 1990s, and the leading indicator qualities of money have diminished. The OECD suggests that the ECB may be therefore be placing too high an emphasis on money in its two pillar-strategy. Furthermore, the strategy poses a communication challenge, as it is unclear to outside observers how the ECB uses the monetary pillar, and how much weight it attaches to it.

 

The OECD is making four precise recommendations, short of getting rid of the money pillar. They are designed to help the ECB manage the monetary pillar more effectively:

 

1. The ECB should publish additional quantified medium-term inflation forecasts, using money-based models. The ECB already publishes forecasts on the basis of the cashless New Keynsian models.
2. The ECB should improve the statistical reporting of monetary trends. At present, it publishes M3 adjusted for portfolio shifts – which recent research suggests may have better leading indicator qualities than M3 proper. The OECD quotes Otmar Issing who said the ECB had development a framework for extracting the signal in monetary developments relevant for policy. The ECB should therefore publish these low-frequency components of M3.
3. The ECB should extend the horizon of its forecasts, as it already uses forward market data.
4. It should try to quantify the risk-distribution around its forecasts, perhaps by using a fan chart or other such tools.

 

 

 

It is relatively safe to predict that the ECB will not heed most, or any these suggestions. The ECB is not using a monetary model of the euro area economy, and would therefore not be in a position to deliver a monetary-based forecast of future inflation. The ECB’s argument in favour of money is not that money is a reliable early indicator of inflation, but that money carries information that needs to be analysed. The OECD’s second point – publication of the low-frequency component of monetary aggregates – is perhaps the strongest of all four. The ECB indeed filters the money data into high and low frequency components, and uses this low frequency time series as input in its policy discussion. But the decision is ultimately a judgement call, which is why in the OECD’s own words, it “poses a communication” challenge. Point four is the weakest, and the one most certain to be ignored. While the fan chart is high illustrative, it is generally considered not to deliver a true picture of underlying risks, which are understood less than the multicoloured chart would suggest.

 

 

Hidden in an annex to the monetary section, the OECD effectively makes another recommendation: to dump the headline rate inflation index in favour of a core index. (for an opposite argument: see the FT column Why they take strawberries out of the basket by Wolfgang Munchau). The OECD focuses on the question whether headline inflation lags core inflations (as the period around the millenium would suggest), or the other way round. The OECD concludes from its own model that there is feedback in both directions, but the extent to which core attracts headline is four times stronger than the link in the other direction. In other words: core is the better indicator.

 

 

How can this be consistent with the statistical evidence at the turn of the millennium, and most recently. In 1998, the oil price fell to $11 pb, and went back up to $30 in 2000. While the headline rate reflected this increase fast, the core rate eventually caught up. But the main reason for the increase in core was the closing output gap, due to the economic recovery at the time, while the main reason for the rise in headline inflation was the oil price. A similar claim could be made for 2006, when core seemed to lagging headline before picking up.

 

 

---