Tracking news over the last year suggests that there is significant imprecision in language from both policymakers and analysts. This summer we learned the imprecision surrounded what was the likelihood of the Taliban overrunning Afghanistan. This fall we found that there was imprecision with the likelihood of key legislation in Congress. With the Fed, there has been imprecision in language concerning tapering and rate increases. There is spin with this imprecision. Policymakers don't want to be pinned to probability estimates. There is also imprecision from analysts who must discuss these issues. The talking heads must provide commentary, yet they do not want to be held to any precise estimates with their thoughts. Therefore, the meaning of words are so important if an analyst truly wants to provide a forecast and add value.
We have talked about Sherman Kent's great work to solve the imprecise language problem at the CIA. He tried to bridge the disconnect between language and probability through forming a precise link between the two. Use common language; however, make sure that the audience knows what the words used mean.
The results graphed above are based on thousands of actual responses to an online survey created by Andrew Mauboussin, a data scientist at Twitter, and his father Michael, research director at BlueMountain Capital Management. The details of the survey can be found in "The perception of probabilistic value for different expressions", and a drag tool can be found in the following article, "Here’s a handy tool to help you talk about probability."
For more on Sherman Kent and words of estimative probability, see:
No comments:
Post a Comment