The complications caused by the absence of a properly representative national house price index have been illustrated again via a speech given by NAMA’s Chairman Frank Daly (see NAMAWinelake here). Frank discusses NAMA’s assessment of the residential sector as follows:
On the residential sector the Central Bank is forecasting falls of 60% from peak (end 2006) to end 2012 under its adverse scenario or 55% under its baseline scenario – based we understand on the PTSB\ESRI index. At NAMA we are not surprised by this and it is not as alarming as one would first think. We do not believe that the PTSB\ESRI index currently showing close to 40% fall from peak is realistic and reflective of where the market is. NAMA’s base valuation date was November 2009 and at this date we were already taking account of on average 50% falls in residential property values from the peak.
So while the residential market may have some little more to fall and no one can be certain that an average fall of 60% from peak may not occur in residential house prices, we would believe that the bulk of this has happened already.
Based on my own anecdotal sample, I’m inclined to agree with Daly that residential prices have fallen more than shown by the PTSB\ESRI index. However, the implications for the Central Bank stress testing exercise strike me as a little more serious than Daly suggests. Daly indicates that most of the peak-to-trough decline envisaged in the Central Bank stress scenarios has already happened.
But this raises the question as to whether the stress scenarios should be based on a peak-to-trough calculations or should they be based on an assumption about a current level of prices and an additional assumption about further declines. It’s not clear why the scenarios should be based on a peak-to-trough assumption. And if, for example, the valuation of residential mortgage portfolios is based on an inaccurate assessment of current levels of house prices, then this may undermine the credibility of the calculations. I would hope that the report accompanying the stress test results would discuss this issue.