Tuesday, December 4, 2012

Hyperinflation: Still Doubtful Based on Market Implied Probabilities

I saw this quote purportedly from shadowstats, but I have not seen the original,

"I assess the chances of a U.S. hyperinflation being underway by the end of 2014 at more than 90%, by the end of 2013 at more than 40%. A likely trigger event here, again, would be panic selling of the U.S. dollar and dumping of dollar-denominated paper assets such as U.S. Treasuries. The initial trigger event could be seen literally at any time."
What's funny/sad about this probability estimate is that you can calculate various market implied probabilities of hyperinflation.

For example, you could buy a January 2015 $2450 gold call for $43.55 and sell a January 2015 $2500 gold call for $40, for a net debit of $3.55 for a call spread which would be worth $50 if gold cost more than $2500 in January 2015. (And surely it would, if there was "hyperinflation" by the end of 2014.) The market implied probability of hyperinflation before January 2015 is $3.55/50 equals 7.1%. His estimate varies (is wrong?) by an order of magnitude.

You could also calculate market implied probabilities of deflation. Buying a $900 gold put for $8.40 and selling an $850 gold put for $6.70 costs a net of $1.70 for the January 2015 put spread which would be worth $50 if gold fell below $850 by then. That gives an implied probability of $1.7/50 which is 3.4%.

So it's interesting that the market thinks hyperinflation is twice as likely as deflation, but that the probability that neither one will happen by January 2015 is 89.5%. (Unless you think gold wouldn't go significantly up or down in inflation/deflation, which seems unlikely because that is the whole point of owning gold.)

We know from cognitive bias research on oerconfidence that when people (most people) give a 90% probability estimate, the observed probability is much lower. They are not well calibrated.

By the way, you know what's a decent protection against inflation? Treasury bills. If nominal risk-free interest rates are equal to the expected inflation rate plus a real rate of return and real returns are somewhat constant, then Treasury bill rates will move closely with inflation rates and behave as an inflation hedge. There is empirical evidence that this has worked.


John said...

"Weatherman statistics", maybe? LOL


CP said...


Pretty bad probability estimate, though.

CP said...