Feb

22

My firm, Manchester Trading, got its start some 40 years ago, based on the idea that interactions between multivariate time series were important. Since then we have proven this fact many times over. As in ecology, 'there is always a food web between the various species in an ecosystem', but our problem is that the web is always changing. There is always an intimate relationship between concurrent correlations of changes of commodity pricing, but the predictive properties are always changing.

In the bad old days of the early 80s it was very common for days like yesterday to occur. A tremendous increase in gold in conjunction with rises in all the grains, and a decline in bonds. Those days it was always good for a major decline in stocks that day or the next. Yesterdays move in gold, up 21.5, in conjunction with a 2% increase in the Goldman futures index and fantastic up moves to multi day highs in all the grains and metals, brought back those ugly memories.

To put things in the simplest perspective, taking a page out of the book of yesteryear, and from localized contingent regression predictions, I found that gold on the day has been up two standard deviations or more on 58 occasions since 1996. The impact on stocks five days later is down half a percent, with no significant effect on bonds, gold, or any other commodity. But still, ugly memories …


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search