Michael Shaoul just made more of my braincells die by saying something stupid about Japan's GDP report.
...this recovery was insufficient to lift aggregate GDP back into positive territory, with the official estimate showing a -0.4% decline in GDP QoQ (-1.6% annualized), up from -1.9% in Q2.
See the problem?
It jumps out more when you're me and you read something even more idiotic on the weekend (probably on some joke site like Business Insider) saying that Japan's Q2 GDP was -7.6% annualized.
You do not annualize GDP growth rates by multiplying by 4 unless you want me to hunt you down and beat you to death with the severed limbs of those you love.
The "annualize" in this case means nothing: all you're doing is multiplying a very noisy, variably trending, dataset by 4.
Yes, sure, if the Q3 -0.4% move in GDP remained exactly the same for 12 months, you'd get an "annualized" -1.6% drop; but Japan's Q2 was -1.9%, and Japan's Q1 was a revised +1.6%, and so far that totals up to -0.7% for the three quarters so far.
See how far away "-0.7%" is from "-1.6%"?
Shaoul's thesis is that Japan will grow just fine, and the recent bump was the result of the sales tax increase which has now worked its way through the system; so if Japan gets back to a (say) +0.4% growth in Q4, that would mean an "annualized" GDP growth in 2014 of -0.075%.
If Japan instead got back in Q4 to just half the growth rate of Q1, that would give an "annualized" growth rate of +0.025%. (Still sucks, but as Kruggers would say that's what you get when you implement the contractionary policy of a sales tax increase.)
Shaoul probably took math in school, so I don't see why the hell he feels the need to "annualize" the butt end of a one-off pothole.
Don't go "annualizing" things that can't be "annualized". In a variably trending, noisy dataset, all you're going to do is confuse everyone with bullshit numbers.