Monday, July 7, 2014

A Loser's Guide to Averaging Losers

I firmly believe that the art of averaging down is the financial market's version of eating fugu fish.  One wrong move and it's all over, yet people will never stop doing it because if done correctly, it is the greatest feeling in the world.  First, my definition of averaging down/averaging losers: you buy stock ABC at $50 because of reasons X, Y, and Z.  Next thing you know ABC is now at $35.  You can sell and admit that you made a mistake.  OR you can buy more at $35, in the unwavering belief that reasons X, Y, and Z are still valid.  From here, anything can happen.  In one universe, ABC goes to $100 and you start printing up business cards with "value investor" on them.  In another, the stock does nothing and you eventually sell at breakeven (or even back at $50) and you breathe a sigh of relief: at least it wasn't at a loss.  Or in the universe where everything goes wrong (aka this universe) that stock ABC is actually Enron or Lehman and you average down your net worth all the way to 0$.  Yes, averaging down for untrained amateurs is a recipe for disaster.  Here is how to do it the right way.

Paul Tudor Jones famously had a piece of paper on the wall of his office that read "Losers average losers."  I liked it so much I had it monogrammed in the inside of my suit. And yet most of my positions are adding to positions already in the red.  Is my subconscious telling me I'm a loser and deserve to eventually fail?  Maybe, but this suboptimal strategy is a goldmine for talentless hacks like me as the professionals never do it and amateurs always do it wrong.

First, averaging down only lends itself to simple portfolios made up of an index fund rather than a stockpicker's portfolio of 6-12 stocks.  Averaging down one stock while leaving another alone messes with the allocation percentage and gets too complicated.  Only time this would work is computing a basket of stocks and averaging down the whole basket.  If you are running a long/short portfolio, forget it, this is getting way too complicated.  Might as well get a real job.

Next, you run into the problem of drastic underperformance when there is no correction.  By definition averaging down requires cash, but any cash in a portfolio creates a drag in performance in bull markets.  The typical solution would be using margin, but margin fees really cuts into returns and compounded with the threat of margin calls (remember averaging down means your positions are red and getting redder) makes a trip to a psychiatrist and an Ambien prescription unavoidable.  The better answer is leveraged etf's.  This is doubly awesome because this is another instrument that professionals sparingly use and amateurs use too much.  No Wall Street portfolio manager will be caught dead holding leverage etf's long enough to bag long term capital gains. Amateurs who do it eventually lose everything in any bear market as a ho-hum 25% drawdown becomes a disastrous 50%+ loss in leveraged etf's.

Here we identify the one odd situation where averaging down works: if you have an edge in identifying broad-market risks.  By understanding macroeconomics and the underlying trends that drive long term (6 months) bull and bear moves, you can use averaging down to achieve superior returns while surviving the bear markets, ironically by hibernating.

Putting it all together, here is how it works (and essentially what I do), first using my voodoo dark black magics macroeconomic analysis, I decide on a confidence percentage: the probability that the overall market (SPY) achieve higher prices in the next six months.  Right now, that number is ~78% thanks to our merciful and magnanimous Fed Chair Janet Yellen (alongside other things).  I then do a quick and dirty beta calculation to figure out how much IWM and TNA to buy to mimic a 78% SPY position (~54% and ~18% respectively).  Now I take the 54%, multiply it by 2 to mimic the margin portion and subtract the 18%.  This gives me the "dip buying" funds.  This means currently I have 90% of my portfolio to buy dips with. 

Actually figuring out when to buy dips is a mechanical affair.  It doesn't matter if you use support/resistance lines, bollinger bands, stochastics, RSI, etc.  As long as you stay consistent and unemotional, they all work equally well in the long run.  Personally, I chose to use a modified Wilder's RSI and have enough reserves to buy dips 15 times. When they rally enough depending on my formulas, I sell it.  Again, I emphasize this is all mechanical.  A robot or even a public school educated individual can do this portion.

Here is what you expect to see with this kind of speculating style.  First, the edge comes from your analysis of the macro, not actually buying dips.  If you think it's a bull market and it's bear, averaging down will lose you a lot of money.  If you think vice versa, have fun looking on from the sidelines while everyone else makes money and start appearing in hedonistic rap videos.  This is why I like dealing with percentages rather than a binary bull/bear market.  As an example, this current bull market is in its 5th year.  I agree it is a bull market, but it is no longer a 100% equity allocation balls to the wall bull market as it was in 2009.  80% is a much more comfortable place to be.  Decreasing this number will depend on seeing more risk factors appear, such as increase in unemployment claims, decrease in ISM, decrease in business/'consumer confidence as well as deteriorating technicals such as bad breadth, distribution on high volume etc. 

There are still a lot of idiosyncratic risks.  Sometimes IWM and SPY diverge for months on end (or even years as in the late 1990s).  More likely, I can just be dead wrong about the macro.  One always has to entertain the banal notion that he isn't as smart as he thinks he is.  Regardless, averaging losers has been much maligned.  Just because most people have been doing it wrong doesn't mean it cannot be used correctly.