Much has been made about the fall of Apple, Inc. (NASDAQ:AAPL) stock price from above $700 to below $450 in a matter of months. I want to use some statistical tools to show that AAPL's price drop has similar features to the deleveraging in the equity markets seen in late 2008/early 2009. Based on my thesis that AAPL's drop was due to liquidation of levered long positions rather than weakening fundamentals, I recommend AAPL investors to buy on weakness.
What statistical tools am I talking about? Well, alpha and beta, of course. I use alpha and beta all the time as a measure of what a stock's real outperfermance is vis a vis the market.
First of all, what are alpha and beta? Many people throw around these terms without actually understanding what they mean. Beta is the ratio that you have to buy/sell a security against another security to achieve a maximum Sharpe ratio (maximum return, minimum variance) portfolio. Alpha is the return of that portfolio, or the outperformance of the one security against the other. For example, consider the following plot of PowerShares QQQ Trust ETF (NASDAQ:QQQ) logarithmic daily returns vs SPDR S&P 500 ETF (NYSEARCA:SPY) logarithmic returns since 2007, adjusted for dividends and splits (all data in this article is taken from Yahoo Finance and manipulated using the statistical language R with the help of the excellent package quantmod):
(The scale is logarithmic, a 0.1% change corresponds roughly to 0.0001 on this plot, and a 1% change corresponds roughly to 0.004 on this plot.)
Beta is the slope of the line running through the ellipse (0.93), which indicates that you would have had to buy $0.93 worth of QQQ for every $1.00 you sold of SPY to achieve the theoretical maximum Sharpe ratio portfolio over this time period.
Alpha is the y-intercept of that line - 0.0002, or an annualized 6% - maybe enough to cover the cost of borrowing to short SPY. In other words, it is a measure indicating the outperformance of QQQ over SPY.
What I want to look at now is the performance of iShares Barclays 20+ Year Treasury Bond ETF (NYSEARCA:TLT), a tracker of the Treasuries market. Unlike QQQ, the price of TLT is negatively correlated with the price of SPY, which means that a theoretical maximum Sharpe ratio portfolio involves going long both SPY and TLT. Here is a plot of the beta (calculated over a rolling 200-day window) of SPY vs. TLT since 2007:
As you probably know, the end of 2008 was a period of huge deleveraging in the markets. And you can see that the beta of SPY vs. TLT spiked heavily during that period, so that instead of SPY moving about -0.4 percentage points for every percent move in TLT as usual, it was moving around -1.4 percentage points. This was because of market participants trying to liquidate their positions due to margin requirements, even at unfavorable prices relative to bond prices.
The spike in beta was accompanied by a similar downwards spike in alpha (again calculated over a rolling 200 day window):
So as an indicator of deleveraging, we can look to an overlay of a spike in the magnitude of beta and a spike downwards in alpha. Here are plots of AAPL's recent behavior in terms of beta and alpha, calculated over 200-day rolling windows.
You can see a spike in beta and alpha similar to SPY/TLT in late 2008. I assume that AAPL's recent move downwards isn't really a reflection of the fundamentals of the company - which are excellent, as discussed in numerous places - but an effect of extremely leveraged long positions getting liquidated.
If you had bought in late 2008 based on fundamentals while everybody else was getting knocked out of the ring by margin requirements, you would have profited handsomely. As for me, I am long AAPL and buying on weakness.