No Curve Fitting
The Macrowonk Indicators have not been derived by looking at market returns and then curve-fitting the indicators to produce the best historical return. This is a very important point. When we first looked at the applicability of the indicators to enhance trading and investment decisions, it was a pleasant surprise to see just how well they perform. Of course, on a logical level, we knew that markets reflect the underlying macroeconomic conditions over the long term (that’s not to say that disconnections don’t happen, but they’re generally over a shorter timeframe). These case studies are designed to provide a small sample of this.
No Look-Ahead Bias
A lot of work has gone in to prevent any sort of look-ahead bias that could be present in the data. For example, we do not use revised economic data, instead we use the original release figure to construct our indicators. The published indicators are as they would have been in history.
No ‘Optimised’ Data Sampling
We have not cherry picked the data window to present a favourable view. All examples cover the period from January 2004 to December 2016 because that’s the timespan of our data set.
On to the case studies:
Notes [ + ]
|1.||↑||Sorry for the clickbait title on this one|