Ask Colin

Yahoo breadth data is different to other sources. Who can I trust?

Breadth data is very difficult to calculate accurately. Some of the problems are:

  1. What to include. Some methods include every listed security. However, this tends to count companies, which have more than one type of security listed, more than once. One approach is to include only the three-letter ASX codes. However, even this is not perfect because some companies do not trade as a three-letter code.
  1. You have to be meticulous with adjustments. If there is any change that requires adjustment of past data, say due to a split or consolidation, you need to be very careful that the adjustment is done before calculating the breadth data for the ex date and subsequent dates. There are also areas of disagreement about adjustments. Different vendors may make different adjustments. Right, wrong and trust are tricky concepts in this area.
  1. You have to be meticulous with name changes. If there is a name change, you need to be careful to join the old file onto the front of the new data file before calculating the breadth data for the ex date and subsequent dates.
  1. Likewise with code changes.
  1. Likewise with calls and instalments that create a fully paid share. Many vendors do not make this adjustment correctly in my opinion.
  1. Back-door listings are a tricky subject. Some of us disregard the old data, but many vendors retain it. This will give different breadth data between vendors.
  1. Demergers are also very tricky. Each one needs to be considered on its merits. There is a lot of room for different approaches to joining or not joining old data to the front of the new file (and which file?) as well as what adjustment to make.

Conclusion

The breadth data will vary from source to source. It is not so much a matter of trust, although some may be more meticulous than others in maintaining their database. Putting aside incompetence, there are genuine and eminently arguable reasons why different vendors will use different methodology. The key I think is to use a consistent source.

Keywords: