A common error: confusing direction, level and intensity

One of the most common mistakes among Wall Street participants — experts included — is 0ver-reaction to a change in direction. This occurs when there is news that shows direction, and direction alone.

Interpreting data requires three different pieces of information:

  • The prior level — how much of the information was already reflected;
  • The direction of the change;
  • The amount of the change, what we call intensity.

An Interesting Example

Paul Kedrosky’s Infectious Greed is a charter member of our featured links. We profit daily from his observations, and especially his work in highlighting excellent research from a wide variety of sources. This is material that would often go unnoticed, since he often catches things that others miss. We read him daily, and so should you.

We are a little surprised at his conclusion about the Fed survey of lending practices, a conclusion not supported by the data. It is a common mistake. Here is the chart of the data cited:

6a00d83451ddb269e200e553fd1ed28834-450wi Paul’s conclusion is as follows:

We are at levels of credit restrictiveness that we haven’t seen in the modern history of the Fed loan officer survey.

This conclusion is not supported by the data, which show a quarterly change in the direction of restrictiveness. We know that many banks are more restrictive than last quarter. Many of them also moved in the same direction in the prior quarters. The data do not tell us the starting level of restrictiveness nor the amount of the change.

Assume, for example, that banks were using a particular FICO score for a certain loan. We have no idea how the current levels compare with those from five, ten, or twenty years ago. There is no way to tell if this is the highest level of restrictiveness — or not. We do not know if many banks made a small change or a massive one. They might still be at liberal levels — or not.

Our Take

Please note that we are not rejecting Paul’s conclusion. The data do not support that either. In fact, we agree with his basic thrust that lending should have been tougher a few years ago. Who would not?

Our point is that data interpretation requires careful thought about what questions were asked and what conclusions can be drawn. Paul’s interpretation was (uncritically) accepted by other mainstream sources.

Real Time Economics has a more careful and accurate interpretation.

In fact, we do not know if this is a slight change by many banks or a major change. We also do not know whether the current levels are much tighter than those in the “pre-bubble” era. It would be more helpful if the Fed survey was anchored somehow, showing the overall level and the degree of change.

This observation has many applications. Market participants often react to changes in economic indicators without a good feel for what is already reflected. The applications include the following:

  • Labor force participation
  • Unemployment rates
  • Earnings changes compared to prior cycles
  • And many others.

Originally published at A Dash of Insight and reproduced here with the author’s permission.