Sunday, June 3, 2012

Gold vs. US debt linear regression update



The R-squared is 0.93.

Best fit price is 1,679.80.
One sigma below is 1,597.07.
One sigma above is 1,762.53.
Two sigma below is 1,512.79.
Two sigma above is 1,846.81.

Usual and customary disclaimers. Correlation is not causation. This is not investment advice.

Charlie Munger says that gold is only useful to Jews in the 1930s.

Warren Buffett says stocks beat gold, but he has been wrong for the last 11 years in a row. Maybe Buffett will be right this year.

6 comments:

  1. Serious question.

    Using your trend line, what would the price of gold be if the debt was zero?

    I'm curious.

    Just something to think about.

    ReplyDelete
  2. Stagflationary Mark,

    This regression only uses data from 2001 to date. Using only that data, the price of gold would be negative if the debt was paid off, both of which are clearly impossible.

    Take a look here, something big changed around 2000 in the US and world economy. Party over.

    If the debt gets a jubilee, or gets worked off after 30 years or so, we might expect gold to go back to it's longer term trend since 1971.

    While this model has proven useful in recent years, it could end at any time. The entire gold thesis I have been sticking with may prove wrong in the end.

    ReplyDelete
  3. This regression only uses data from 2001 to date. Using only that data, the price of gold would be negative if the debt was paid off, both of which are clearly impossible.

    That's pretty much my point.

    Take a look here, something big changed around 2000 in the US and world economy. Party over.

    Using my own charts against me! Oh the humanity! ;)

    While this model has proven useful in recent years, it could end at any time.

    I guess that's my point. Beware the "sure things". I'm especially skeptical when a trend doesn't extrapolate backward in time all that well. And yes, I have done similar trend lines myself. Just pointing out the risks. There have been so many trend failures in recent years that I don't really trust any of them (my own included).

    I'm still curious what the price would be. How negative?

    ReplyDelete
  4. Using just the data for this model, if US debt went to 0, the gold price would be -628.49. Both conditions seem unlikely.

    I am trying to contemplate how the the US debt could reach 0. It could happen by Congressional action defaulting on 100% of the debt or extreme hyperinflation. Neither outcome would be pleasant for anyone.

    ReplyDelete