Working Up

by Dwayne Phillips

Linear Regression is an old, basic technique to predict the future given the past. It seems that its utility has been forgotten by many.

How did we get here? Why didn’t anyone see this coming?

These are a couple questions I heard recently when groups of persons gathered to discuss … well, big booboos (mistakes) in finance. It seems that incoming funds did not match what was predicted; expenses were what was predicted, and, uh, well, the money ran out.

Why didn’t anyone see this coming? Seems someone forgot about linear regression. This is a technique created in the first decade of the 1800s. You plot a few data points from early in an endeavor. You draw a line that fits those points. You look where that line goes out into the future. The formula for this looks daunting, but calculating it is pretty easy. Excel and other spreadsheets have an easy-to-use function that does this for us. Viola’. Done.

Of course there are many cases in which this simple prediction fails. Things happen. People happen. Times change. Hence, some thought must be applied. Still, this little technique does predict the future pretty well.

Why is it not used? The biggest reason I have observed is that people want to believe that things will work out. Bad news is not welcome. Bad news simply must be incorrect. Things will change for the better.

Question: If we have done such-and-such eight weeks in a row, why will we do differently in the future?

Answer: Well, there are many possible good answers. Those answers require us to change. We don’t like to change and we don’t know how to create change.

Hence, this little, old, simple tool works most of the time. How did we forget about it?