With SPY recently surpassing the $100bn asset threshold, it is time to re-visit an over-looked counterpart to the capitalization weighted SPY: equal weighted RSP.
RSP consists of the same 500 stocks as the S&P 500 index, but instead of weighting the holdings by their market capitalization, it weights each holding equally. Instead of the S&P 500's largest component, Exxon Mobil (XOM), contributing 3.50% (as of 1/20/2012) of the fund, and smallest component Titanium Metals (TIE) contributing 0.11% of the fund, each holding is weighted at 0.2%.
RSP, the Rydex S&P Equal Weight ETF, has been in existence since only April 30, 2003, ten years less than the first ETF, SPY. However, RSP's benchmark, the S&P Equal Weighted Index, has been tracked by the CME since the end of 1989 giving us a longer horizon for comparability between the two funds. RSP has held up well against its cap-weighted big brother.
The 209bp of annual outperformance is in part a function of RSP's equal weighted nature, which gives it a larger weight to mid-cap stocks versus blue-chips. The risk, however appears justified when you compare the Sharpe ratio of the two funds. (The average yield of the 10-yr Treasury over the sample period was used as the risk-free rate.)
Even factoring in the higher expense ratio of RSP (40bp) for the full dataset versus the ultra-low expense ratio of SPY (9.5bp) yields a higher Sharpe ratio after fees for RSP.
We now know that RSP, over an extended time horizon, has provided superior alpha to investors versus SPY, but specifically what is this risk? It is important to be able to categorize the specific risk premium attributable to RSP in order to properly weight your portfolio allocations. Certainly, the equal weighting gives RSP a higher allocation to smaller capitalization companies, so we regress its returns versus the S&P 500, the S&P Midcap 400, and the S&P Small Cap 600. The underlying index for RSP and the S&P 500 index have a correlation coefficient of 0.89. The underlying index for RSP and the S&P Midcap 400 index (available through MDY with a 25bp expense ratio) have an even higher correlation coefficient of 0.96. The underlying index for RSP and the S&P Smallcap 600 index (available through IJR with a 20bp expense ratio) had a much weaker correlation coefficient of 0.53. Equally-weighting the S&P 500 certainly gives you a diversified mid-cap type risk. Going down the size spectrum over this sample period has been increasingly positive for returns.
Part of the excess return is therefore attributable to the size function, but could part of the return also be a function of value versus growth? A capitalization-weighted fund, such as SPY, is inherently giving more credence to stocks that have performed retrospectively. Overvaluing these companies relatively could be part of the reason for RSP's relative alpha generation. To determine if RSP is relatively weighted towards growth or value, its returns were regressed against the S&P 500 Value Index (available through IVE with an 18bp expense ratio) and the S&P 500 Growth Index (available through IVW with a matching 18bp expense ratio). RSP is certainly more correlated with the value index relative to the S&P 500's growth component, but this did fact did not positively factor into its relative outperformance of SPY on average.
A recent Bloomberg article highlighted the substantial return differential of an equal-weighted strategy and a capitalization-weighted strategy through the 2000s. If you are a passive indexer looking for alternatives to the S&P 500 and able to take more mid-cap value risk, it may make since to examine RSP as an attractive alternative.
Disclosure: I am long SPY.
No comments:
Post a Comment