Association is not causation

admin — Sun, 12 Jul 2009 12:57:08 +0000

In economics, identifying cause-and-effect relationships is very important. But statistical association alone cannot establish this causation. Perhaps an extreme example will illustrate the point. Suppose that each November a witch doctor performs a voodoo dance designed to summon the gods of winter, and that soon after the dance is performed, the weather in fact begins to turn cold. The witch doctor’s dance is associated with the arrival of winter, meaning that the two events appear to have happened in conjunction with one another. But is this really evidence that the witch doctor’s dance actually caused the arrival of winter? Most of us would answer no, even though the two events seemed to happen in conjunction with one another.
Those who argue that a causal relationship exists simply because of the presence of statistical association are committing a logical fallacy known as the post hoc propter ergo hoc fallacy. Sound economics warns against this potential source of error.

Time Cycles: Four Days to Four Years

admin — Thu, 25 Jun 2009 20:22:07 +0000

Although they are not always apparent, there are a number of fairly regular wave patterns in the movements of stock prices that appear to be based upon time. There seem to be regular and repetitive cyclically determined time periods between low points, which is how cycles are defined. Frequently there are equal lengths of time from highs to highs as well within full market cycles, generally during neutral market periods, but cycle lengths are normally measured from lows to lows of market cycles.

Money, trading, forex, insurance, loans » patterns

Association is not causation

Time Cycles: Four Days to Four Years