Zero Correlation and Independence: Why "No Relation" Isn't Always Simple
Imagine you're tracking two things: the number of ice creams sold at a shop, and the number of umbrellas sold at the same shop. On a hot summer day, ice cream sales go up — but umbrella sales might go down. That's a negative correlation. On a rainy day, umbrella sales go up and ice cream sales go down — again, a correlation.
Now imagine you track the number of ice creams sold and the number of history textbooks sold. Do they move together? Probably not. On a hot day, ice cream sales rise but textbook sales stay the same. On a rainy day, ice cream sales fall but textbook sales don't budge. These two variables seem to have no relationship — they move independently of each other.
That's the everyday intuition behind zero correlation: two variables that don't move together in any systematic way.
The Precise Meaning
In statistics, correlation (usually measured by the Pearson correlation coefficient, r) tells you the strength and direction of a linear relationship between two variables. The value of r always lies between −1 and +1:
- r=+1: perfect positive linear relationship (as one goes up, the other goes up proportionally)
- r=−1: perfect negative linear relationship (as one goes up, the other goes down proportionally)
- r=0: zero correlation — no linear relationship
When r=0, the scatter plot of the two variables looks like a random cloud. There is no straight line that meaningfully describes how one variable changes with the other.
Zero correlation does NOT mean independence. This is the single most important point to remember. Independence is a much stronger condition.
The Critical Distinction: Zero Correlation vs. Independence
Here is where the concept gets subtle and where most students slip up.
Zero correlation (r=0) means there is no linear relationship. But two variables could still be related in a non-linear way. For example, consider the relationship y=x2 for values of x symmetrically distributed around zero (like −3,−2,−1,0,1,2,3). As x moves away from zero in either direction, y increases. The correlation coefficient r will be close to zero — because the best-fit straight line through that U-shaped scatter plot is nearly flat. But y clearly depends on x; they are not independent.
Independence means knowing the value of one variable tells you nothing about the value of the other. If two variables are independent, then their joint probability distribution factors into the product of their individual distributions. Independence implies zero correlation, but zero correlation does not imply independence.
A common exam mistake: stating that r=0 means the variables are unrelated. They are only unrelated in a linear sense. Always check for non-linear patterns in a scatter plot before concluding independence.
Why This Matters in Economics
In economics, you constantly deal with relationships between variables: price and quantity demanded, income and consumption, interest rates and investment. When you compute a correlation coefficient and get a value near zero, you cannot simply declare the variables independent. …