So. A gallery of SupADF tests. Who doesnt like some right tailed unit roots of a saturday morning eh? I have blogged on these tests for bubbles before : see here, here, here . I leave you to make your own conclusions…
Is there a bubble in the Dublin house market? The latest figures might well give one pause for thought, with a 22% year on year increase, clearly unsustainable.
However, this may not be the entire story.
A newly developed statistical technique (see here and here for previous posts using it to talk about bubbles in London and Bitcoin) has, it is claimed, good power to detect bubbles. Below see the results for real dublin house prices, 1996 prices, for new and second hand homes.
We can interpret these as a bubble being present in the series when the blue line exceeds the red line. It clearly picks up the last bubble period, and interestingly dates its beginning to the 1997-1998 period. What it also shows is that there seems to be no such bubble at present.
This is only a statistical test – it is by no means definitive. What is clear is that bubble or no, Dublin prices in nominal terms are rising unsustainably fast.
What is your thought reader?
Bitcoin is soaring. I have previously noted a new test (The generalized sup-ADF test of Phillips for the technically minded : see https://editorialexpress.com/cgi-bin/conference/download.cgi?db_name=CEF2010&paper_id=69 if interested) for bubbles. Applied here we see not one but two bubbles in bitcoin. The shaed series is the test statistic, the light green line its critical value – when the test statistic is above the critical value, it suggests a bubble. Also imposed are the closing values of bitcoin v $US (from MtGeox exchange).
I leave the (pretty obvious) interpretation to yourselves…