Correlation or Causation?

posted by Decision Innovation Solutions on Tuesday, May 15, 2012

I recently had a discussion with an individual about what drives the prices of commodities, which was a good discussion, but it helped remind me about why I enjoy and get frustrated with economics and correlations. It also reminded me of the importance of asking better questions, getting better answers, and making better decisions, which is our niche as a business.

Our discussion turned to exports, land availability, and the value of the dollar. I had to make the disclaimer to the person I was talking with that just because two things are highly correlated does not denote causality; the change in one does not necessarily cause a change in the other. What do I mean by this? It means that, in general, when the value of the dollar is low, exports are up and when the value of the dollar is high, exports will be down.

Another example of this distinction is when exports are up, commodity prices are higher than when exports are lower. There is a correlation that exists between the level of exports and the price of commodities but one does not necessarily cause the other. There might be other underlying principles that are not seen and may not be accounted for. Some of those things might be production levels, how much commodity is in storage, drought, timing of planting and harvesting, turmoil between countries – my list could go on but I will stop there.

To make this principle a little simpler, a study might show that there is a correlation between the number of police officers and number of arrests made. This correlation is a positive correlation meaning that as the number of police officers increase, the number of arrests also increases. The real question is, does the size of a police force determine the number of arrests made, or does the high/low crime area determine the size of police force needed. A person could make a case either way, but remembering that correlation does not denote causality could help us take a step back and make sure that we have all of the data to make an accurate assessment. There are likely underlying factors that might not have a direct correlation to either the number of arrests or the number of police officers that have a potential impact on both of the aforementioned factors.

As a business, our purpose is to "Provide expertise to create effective support tools and strategic decision alternatives." In fulfilling this purpose, we need to remember that correlation does not imply causation. We need to be asking better questions, to fully understand the correlations that we see, and to determine what the underlying principles are.

  1. causation
  2. correlation
  3. dis

About The Author

Decision Innovation Solutions