Regressions in hard commodities
by L. R. James
The possibilities for precious metals price rises are present, but so are the possibilities for their decline.
The bond market and the Federal Reserve give every indication of being intolerant of anything more than
low level inflation, raising interest rates in response. During the past few months, rates have moved
further and further in the direction of an inverted yield curve. A few of the Commerce Department's
economic indicators also have suggested some slowdown in the economy. So far, in this business cycle,
precious metals have not kept up with stocks.
I use stepwise multiple regression to measure the relationship between different variables, like inflation
and gold prices. The process determines a linear equation with which you can estimate the price of gold
for a given level of inflation. Given two stochastically dependent, random variables, stepwise multiple
regression will predict the mean expectation of one variable relative to the other. (I don't recommend
learning regression analysis from a textbook. It would be better to take a course in it. Even easier would
be buying software to do the work for you. StatPackets [$25 from Walnuck Associates, (612) 866-9022]
is a great package, dirt cheap and uses Lotus 1-2-3 or Symphony files.)