Tuesday, April 13, 2010

Liquid Assets? You've Been Hitting the Bottle A Little Too Hard

Today must be a light news day because too much attention has been paid to this ridiculous study of wine price appreciation.  The authors claim investing in wine beats investing in stocks, and aids in diversification of investors' portfolios, reducing risk and increasing returns.  I appreciate a good wine.  I even appreciate a good study.  This paper falls in neither category.  Put this back with the Maneschevitz and Tarot cards.

They examine several years of wine auction trades on a subset of relatively more actively traded wines that one might construe as something investable.


First problem: They construct a data set which imposes their results on the data.  Always read carefully the data construction methodology.  Step five of their process says that when multiple trades in a single wine occur in a month, they take the median price.  Moreover, if a price swings by 40% or more month to month, they throw out the data.  These two steps sound innocuous.  However, they materially impact the data.  Step one reduces the volatility of each and every component of the index.  Step two throws out the riskiest part of the index.

Thought experiment:  Take the Russell 2000 index.  Recalculate using returns to monthly median prices rather than month end prices.  Then, throw out all the companies with 40% month to month price swings.  What do you have?  It certainly isn't the Russell 2000.  It isn't even close to a tradeable index.  I call my Goldman equity derivatives sales guy to buy that index, he's going to laugh, call his pricing guy, who will proceed to rip my face off.  Yes, that is a technical term.

Second problem: They seem to ignore transaction costs.  I know this because the word "transaction" appears twice in the paper.  Similarly, "cost" appears three times.  All five observations of these words appear in descriptions of the data.  I'd venture a guess that 10% commissions may underestimate costs. 

Third problem: Carrying costs.  Back to word searching.  No appearances of the words: "storage", "insurance", "cellar".  How much does it cost you to store and insure your Russell 2000 portfolio?  You don't know?  That's because it's very close to "free".  [Nothing is free, I know.  Custody costs have been buried by your broker in the transaction costs, your lousy return on cash in their money market fund, and the fact that they lend out the securities to others, and you hold their credit risk...I could go on forever here...]  Storing wine ain't free.  And, the insurance burns 2%...every year!

The problems I raise hardly require such deep insight.  However, they make the results irrelevant to anyone making investment decisions. The authors should be ashamed to have their names on this paper as written.  (And, that's before we even consider their methodology and results!)

No comments:

Post a Comment