Just as an aside here, one would expect an artificial increase this year because of the change in counting method to the 15 minute threshhold, which took place in the middle of a service year.
Thus, last year had the benefits of approx. six months of the new counting method, and showed an increase. This year had the benefit of a full 12 months of the new method, and reasonably should also show an increase becasue of it.
Which means that in places like Canada which recorded a big fat 0, the result is very bad indeed.
Next year of course, the stats are likely to return to the long term trends i.e. stagnation. Unless they cook up another statistical dodge.
Finally, changing the method of counting and not highlighting the change in a year to year comparison is incompetent at best, and blatantly dishonest at worst. And in economics, a company changing its methods to make its results look rosier is a classic sign that the company is in trouble.
Expatbrit