Irrational Bond Markets
This post was written by Colm McCarthy
The post hoc attribution of market movements to specific events is always a bit speculative. But Thursday’s release of weak Q2 national accounts (Economy Shrinks Shock!) was headlined in Irish and international media and appears to have unsettled the secondary market in Irish government bonds. The 10-year closed around 6.65 on Friday, the worst close of the crisis to date and cue panic stations.
This makes sense if
(i) The prospective growth rate in 2011 and later is critical for fiscal sustanability, which is reasonable, and
(ii) A sensible person would have revised their expectations on the basis of Thursday’s release, which I think is not reasonable at all.
So far as I know, there have been just three technical studies of the Irish quarterly data since the CSO commenced publication in 1997 and which are relevant to this discussion.
In this paper, the conclusion here is that the Irish quarterly GDP and GNP numbers are very volatile, and considerably more so than is the case in other OECD countries, including smaller ones.
This study concludes that the seasonal adjustment procedure used by CSO is (probably) not the best, and that this can make a difference, in the sense that an alternative, and preferable, procedure can give results which sometimes alter the qoq % changes quite a lot.
The final study shows that revisions to the first-shot estimates, while no greater than elsewhere, can be substantial.
The economy sank like a stone through 2008 and 2009. The last three qoq % changes in real GNP, using the CSO’s seasonal adjustment, were -1.9, then -1.2 and just -0.3 for the most recent Q2 number. Using the alternative (indirect) seasonal adjustment, the most recent number was -0.1. A reasonable headline in either case would have been ‘Economy Now Flat’.
The reason for the Shrinking! headlines was the GDP numbers. The last three qoq changes read -2.5, +2.2 and -1.2. These are the qoq changes, not annualised. At seasonally adjusted annual rates (saar) the decline in real output in Q4 was -9.6%. It then grew at a saar of +9.1% in Q1 of 2010, relapsing to a saar of -4.7 in Q2. If you really believe this, maybe you could make it as a bond dealer.
The gyrations in the quarterly numbers are just not credible. Various people including Garret FitzGerald and Robbie Kelleher have speculated that real output measurement in the MNC sector is mainly responsible for the extreme volatility. There may be a bit of informal smoothing practiced by other countries too.
Forecasts of GDP growth for 2011 seem to be mostly in the 2% to 3% range. My point is not that these forecasts are likely or plausible. The point is this: given what we know about the behaviour of the Irish quarterly macro aggregates, whatever your figure was on Wednesday, there was no good reason for changing it on Thursday morning.