Carol Alexander and Dimitris Korovilas, both from ICMA Centre published a series of excellent research papers on VIX ETNs:
Diversification of Equity with VIX Futures: Personal Views and Skewness Preference, and
Understanding ETNs on VIX Futures. Their latest
Volatility Exchange-Traded Notes: Curse or Cure? continues on the topic.
Abstract:
This paper investigates the trading, hedging and performance characteristics of VIX futures exchange-traded notes (ETNs) and discusses their pros and cons from the perspectives of the regulator, the issuer, and the non-speculative investor. Is this direct trading of volatility in very high volumes a major new source of issuer risk and systemic risk? Or – as advertised by the issuers – do these notes provide a unique source of diversification that should be welcomed by investors and regulators alike? To answer these questions we replicate the ETNs daily indicative values from March 2004 – March 2012, showing that the amplitude and frequency of volatility cycles have indeed increased markedly since the introduction of VIX futures ETNs. On the other hand, we explain how long-term investors can build simple ETN portfolios with uniquely attractive performance and diversification characteristics – provided they hold inverse rather than direct short-term tracker ETNs. Then we focus on the ETN issuers’ hedging activities, where the terms and conditions of early redemption provide transparent front-running opportunities for speculators which in turn increases both the hedging error and the volatility of VIX futures. Furthermore the one-day notice period for early redemption presents a moral hazard problem for the issuer.
My notes: In the section 4.2 Correlation Analysis and its Implications for Roll-Yield Arbitrage Trades the authors show a concrete example of using PCA on vol indexes to create calendar arbitrage portfolios. The first PC explains 95% of variance, which explains the almost-parallel shifts in the term structure, with second component explaining the tilt (3.23% of variance), and third PC explaining the curvature (0.91%) They note that XVIX is historically optimal in return/risk sense, same conclusion that I blogged about when XVIX just launched. Unfortunately XIVX
disappointed in its performance.
Perhaps a more interesting application of PCA here would be statistical arbitrage between VIX futures of different maturities, or constructing a combined portfolio with different volatility futures like VXEM, VXEW, or soon to be launched VXN (Nasdaq volatility index) .
