ECB’s potential sophistry in assessing deflation
Deflation is not yet here – at least not for the entire Eurozone. Inflation is merely very low with a 0.5% year-on-year report in March and several countries (including Spain, Greece and Portugal) already posting negative consumer price inflation. So far the European Central Bank (ECB) has kept its cool. Despite more “dovish” communication, the ECB has stuck to its guns in its last two meetings and not adopted a more expansive monetary stance.
According to the ECB, there are good reasons for that. First, the exceptionally low inflation registered in March is likely to reverse in April due to the simple technical fact that Easter holidays – a source of upward consumer price pressure – occurred in March last year and still lie ahead of us this year. So let’s wait and see.
Second, the ECB is not concerned by the current environment but by evolution over the next couple of quarters or even years. According to the March staff projections, inflation is forecast to average again 1.5% by 2016 (with the caveat of a rather broad uncertainty band ranging from 0.7%–2.3%). Finally, the ECB is also relying on inflation expectations, which should be “firmly anchored” around the target range (an inflation rate slightly under 2%).
This analysis suffers some flaws. Arguing with projections suggests they are more right than wrong, which will seldom be the case, even given broad uncertainty bands. The absolute mean forecasting error for two-years-ahead inflation was roughly 0.7 percentage points. The largest two-years-ahead forecast error occurred in December 2007, when the 2009 inflation rate was seen as 1.8% and the actual value was 0.3%. Even worse, the June 2008 forecast for inflation in 2009 was pushed up to 2.6%, some 2.3% higher than the resulting actual value, explaining why the ECB decided to hike interest rates just before Lehman Brothers went bankrupt.
Ultimately, with this framework of focusing on two-years-ahead projections that show forecast inflation at the targeted level, the ECB can argue for years why it is unconcerned by actual levels of deflation, even though the Eurozone might actually remain in deflation for years. Interestingly enough, usually two-years-ahead projections are given for the first time in December, i.e. in December 2013 we got the first ECB forecast for 2015. The March 2014 projections were the first to venture beyond the two-years-ahead future, by arguing that by the end of 2016, the Eurozone inflation rate would be back on target – as if, without explicitly acknowledging it, the ECB would significantly miss its own target on the downside for the next two years at least.
As for the “firmly anchored” inflation expectations, until last fall the ECB was mainly focusing on the two-years-ahead survey of “professional forecasters.” Since then, maybe due to the fact that those two-years-ahead inflation expectations were starting to drop and were hence not so firmly anchored anymore, the focus has shifted to the five-years-ahead survey of inflation.
There are two issues with this. First, professional forecasters – usually economists – are surveyed. They neither necessarily reflect market expectations, nor the expectations of the broader public (companies and households). They are more likely to abide the consensus of their profession and hence be prone to “groupthink.”
Second, the five-years-ahead forecast of “professionals” is firmly anchored per construction. Any professional forecaster will take the target of the ECB as an anchor point to state a five-year-ahead forecast. In fact, since the survey began in 1999, the five-year-ahead forecast of Eurozone inflation has barely moved in a range between 1.8 and 2.0%, so there is no information content in this indicator.
Death by asphyxia – lack of oxygen – will almost go unnoticed if it happens slowly. One may even experience a dizzy euphoria at first. We believe this is a good metaphor for the onset of deflation. Let’s just hope it doesn’t apply to the Eurozone, and that the ECB’s arguments are more than mere sophistry.