In the case of residents, it grew by more than two hundred percent; in the case of non-residents, the number stood above two thousand percent. What do these numbers tell us about the housing market? Are we about to enter another booming period? Does this mean the changes in policy earlier in the year are producing results?
No matter how extraordinary some figures appear, isolated numbers are virtually meaningless. They need context; they must be judged against trends and typical patterns, where they exist. Then, we need plausible explanatory relationships with other variables that can be measured and tested. In both counts, the figures above are found wanting.
This is not a well-behaved market. It is known to display bursts of activity followed by more or less extended periods of relative softness. It is highly sensitive to the ebbs and flows in the supply of new housing units, and to changes in the conditions affecting external demand. Homologous comparisons can be subject to broad and irregular changes over time.
And that is what we find in the data for the homologous growth rates of this type of credit. For the last full civil years and up to last September, the instant picture is one of great variability. The data points display a very significant level of dispersion. Without getting too technical, the standard deviation is in both case – residents and non-residents – more than three times the average. As a rule of thumb, a ratio over one signals a highly disperses sample. Oscillations are very significant, and extreme figures appear occasionally. In the case of non-residents, the data set shows figures ranging from almost obliteration (minus 97 percent) to stratospheric rise (nearly three thousand percent rise.)
So the data is typically ‘unstable.’ And without further information on the profile of both buyers and properties involved in the recent burst, little can also be said about policy impacts or otherwise. Recent evolution may signal the beginning of a new upswing. But what September figures may be saying time only can tell.
