Hi Stephen,
What you term “natural variance” is certainly a possibility – that’s why I noted the difference between mean and median. However, we need to be clear what we mean by variance: given these are comparisons between prices and estimated prices from hedonic regressions, the variance would have to be a step further up the chain, in particular would need to come from unobserved parts of the property. These are certainly a factor (typically about 30% of the price of the house) but I don’t think that means it invalidates the conclusions, particularly given the tentative nature they were phrased.
I hope to pool all the auction results and include variables for each of the auctions in a regression setting. That should hopefully give us some indication of how likely these results would be to come up by chance. Particularly given the point made above about just how significant each additional percentage fall from the peak is in real terms.
Thanks for the comment,
Ronan.
Rich
,
Dead cat bounce comes to mind, between increasing unemployment, 100,000 mortgages 90 days in arrears, the extinction of the new mortgage – ~3000 in Q1 – 96% down on peak, euro instability growing by the day, a €16 bn overspend in the annual budget to be corrected, the consequential affect on public salaries that will have, etc
In brief very houses are for sale as no houses are being repossessed nor any attempt made to liquidate assets of of those defaulting on loan repayments. 1 in 3 BTL mortgages are in default……please explain the dynamic supporting this statistical anomaly?
Stephen ,
You’re taking 4 auctions with roughly 60 properties in each and breaking them out between Dublin, ROC etc.
I can’t understand how you can deduce anything from that. Surely any movements with this few data points could be just natural variance?
Ronan Lyons ,
Hi Stephen,
What you term “natural variance” is certainly a possibility – that’s why I noted the difference between mean and median. However, we need to be clear what we mean by variance: given these are comparisons between prices and estimated prices from hedonic regressions, the variance would have to be a step further up the chain, in particular would need to come from unobserved parts of the property. These are certainly a factor (typically about 30% of the price of the house) but I don’t think that means it invalidates the conclusions, particularly given the tentative nature they were phrased.
I hope to pool all the auction results and include variables for each of the auctions in a regression setting. That should hopefully give us some indication of how likely these results would be to come up by chance. Particularly given the point made above about just how significant each additional percentage fall from the peak is in real terms.
Thanks for the comment,
Ronan.
Rich ,
Dead cat bounce comes to mind, between increasing unemployment, 100,000 mortgages 90 days in arrears, the extinction of the new mortgage – ~3000 in Q1 – 96% down on peak, euro instability growing by the day, a €16 bn overspend in the annual budget to be corrected, the consequential affect on public salaries that will have, etc
In brief very houses are for sale as no houses are being repossessed nor any attempt made to liquidate assets of of those defaulting on loan repayments. 1 in 3 BTL mortgages are in default……please explain the dynamic supporting this statistical anomaly?