Last month I published a graph that I received from a SEFBlog-reader and contributor, see 1: it was about a new leading indicator for the Euro Area. Here's the up-dated figure: it confirms last month's outlook i.e. the (well-known) `Coin is flat. The current downturn in the Euro Area has reached its minimum: from now on the negative pressure on the economy should continuously relax (conditional on survival/viability of the entity called `Euro Area').
Blue line: new fast leading indicator emphasizing monthly growth.
Black: Eurocoin (I guess it emphasizes quarterly growth but I'm not sure about this...)
Green: GDP Euro Area (yearly growth).
Red: don't know...
Filter parameters of the new indicator are not up-dated: performances are truly out-of-sample.
I'm discussing rules/tricks for effective filter up-date with the author of the above indicator. Let's see if my tricks will be useful...
I assume that SEFBlog-readers are aware of the fact that I don't up-date parameters of filters with good real-time characteristics. Which is the very purpose/aim/goal of `customization'.
The turning-point of the new fast indicator was first detected last month, see 1. It is now confirmed and the Eurocoin (black) is flat, too. I expect the `Coin to confirm the turning-point next month.
Two Hours Later: a Few Clarifications by the Author of the Indicator
I received a very useful feedback by the author of the indicator:
There ub=pi/9; so, to get the turning point, I used Lag<0.
This time, the blue line is a result of Lag=0, ub=pi/7.
Blue line is a result of targeting interpolated q-o-q target divided by 3; the output is then multiplied back by 3, so, basically, it tracks q-o-q growth.
Black: Eurocoin that tracks q-o-q growth
Green: interpolated q-o-q growth of EA GDP
Red: a finite symmetric filter (2-yrs long only) applied to green line, a benchmark
Filter parameters for the blue line are not updated since 2011June."
Sounds technical, isn't it?
Whoever designed an economic indicator - a real-time indicator - knows how crucially important design-information is: good (a priori) design-decisions are the key to outperform (outperform those who made wrong(er) decisions...).