Hard to find non-correlation
Traders have certainly noticed that stock prices have gotten increasingly correlated in the past two years. In this highly bipolar market, pretty much all stocks go up when “risk is on” and all stocks go down when “risk is off.”
My study of the behavior of S&P 500 confirms that correlation is running at record levels. The average trailing two-year correlation between S&P 500 stocks soared to about 60% in the fall. This is higher than during the burst of the Internet bubble or the 1987 stock market crash.
Correlation has a devastating effect on our portfolios. Remember that the variance of a portfolio is made of two parts: (1) the individual variance of the portfolio’s components and (2) the product of the cross-correlations of its components.
Higher correlation mathematically leads to more volatile portfolios, which explains why the VIX (NYSE:VXX) doubled from last year’s levels. Higher volatility means that protecting your portfolio with options gets more expansive.
Higher correlation leads investors to pile into the few remaining diversifying assets like the IEF and TBT ETFs and push their prices to absurd levels. The 10-year U.S. Treasury yield hovers around 2% and gold is back above $1,700 an ounce – even though the bullion got caught up in the correlation trade lately.
Higher correlation also hurts the few managers that still engage in fundamental research: what is the point of forecasting earnings and understanding businesses if all stocks, good or bad, trade as one? Long-short strategies fail to make money as managers are simultaneously buying and selling beta.
Who is to blame for this situation? Pundits usually point to the European debt crisis, which allegedly cause markets to become “more macro-driven”. I find this argument extremely weak. “Macro events” and “change in sentiment” can be used to explain any market pattern ex-post. I do not see why the European debt crisis is more macro than the 9-11 attacks or the 1998 East-Asian financial crisis.
The growth of exchange-traded funds is a much better explanation for the spike in correlation. Correlation increases because investors sell and buy large baskets of stocks simultaneously. This is exactly what ETFs do. Contrary to mutual funds, which can pick which stock they sell when they face redemptions, ETFs have to sell the entire basket. Large ETF inflows and outflows can have significant effects on the prices of small-cap stocks, even in the absence of any change in the companies’ business fundamentals. The correlation bubble matches the explosive growth of ETF assets, which recently passed the $1.5 trillion asset mark.
Despite this criticism, I remain a deep believer in ETFs: ETF providers such as Vanguard and Blackrock (NYSE:BLK) greatly benefited retail investors by giving them access to most asset classes at almost no cost. Low-cost ETFs are also the best protection against the supposed “benefits” of actively-managed mutual funds.
The growth of the ETF market has been so quick that markets did not adjust in time. If markets were perfectly efficient, deep-pocketed long-term investors would capitalize on the correlation inefficiency. But markets are not perfectly efficient, investors prefer to play the momentum game rather than wait for the correlation bubble to burst and the ‘correlation of 1.0' keeps hurting everybody’s portfolios.
No comments:
Post a Comment